Top MCP Servers for Marketing Mix Modeling (MMM) and Incrementality Testing (2026 Q3)
Research disclosure: Sellforte published this research and is one of the eight vendors assessed. Claude and ChatGPT assigned the scores from public documentation reviewed on September 23, 2026. See limitations and disclosures.
Quick Summary: Top MCP Servers for MMM and Incrementality Testing
What is the best MCP server for MMM?
Sellforte MCP received the highest total research score in this comparison of eight vendor MCP servers for Marketing Mix Modeling (MMM) and incrementality testing. The ranking by total score is Sellforte (38.3 out of 48), followed by Recast (31.4), then Triple Whale (30.6).
Sellforte had the highest score in five of the eight categories: marketing data reporting, historical performance insights and causal explanation, campaign and ad set-level optimization, MCP interoperability and workflows, and enterprise-grade platform. Recast had the highest score for channel-level optimization and incrementality testing, and tied with Triple Whale for analytical backbone.
What is the best MCP server for incrementality testing?
For incrementality testing through MCP, the highest research scores went to Recast (3.6 out of 5), followed by Measured (3.0), then Sellforte and Haus (2.6 each). On the three criteria for reporting test results (geo tests, own-media A/B tests and Meta Conversion Lift tests), Recast and Sellforte received the highest combined scores. Measured received the only full score for prioritizing experiments. The category also covers designing experiments.
Which vendors offer an MCP server for MMM and incrementality testing?
This research evaluated eight vendors. In order of total research score out of 48, they are Sellforte (38.3), Recast (31.4), Triple Whale (30.6), Lifesight (29.8), Measured (26.6), Fospha (24.3), Haus (18.9) and Northbeam (17.6). Teams that run their own model have separate options, covered in open-source and self-built MCP servers for MMM.
Research scores by vendor and category
Scores are the average of the Claude and ChatGPT evaluations. Blue shading indicates the share of available points; green outlines mark the highest scores, including ties.
| Research category | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam |
|---|---|---|---|---|---|---|---|---|
| Total research score / 48 | 38.3 | 31.4 | 30.6 | 29.8 | 26.6 | 24.3 | 18.9 | 17.6 |
| 1. Marketing Data Reporting via MCP / 6 | 6.0 | 5.0 | 5.3 | 4.1 | 4.9 | 4.3 | 3.1 | 5.0 |
| 2. Historical Performance Insights & Causal Explanation via MCP / 5 | 5.0 | 4.1 | 3.6 | 3.8 | 3.6 | 2.4 | 1.9 | 1.6 |
| 3. Channel-Level Optimization with MCP / 8 | 7.4 | 7.9 | 6.5 | 5.6 | 4.5 | 4.6 | 2.9 | 2.3 |
| 4. Campaign & Ad Set-Level Optimization with MCP / 8 | 5.0 | 1.0 | 3.8 | 4.1 | 2.4 | 3.3 | 3.1 | 1.4 |
| 5. Incrementality Testing with MCP / 5 | 2.6 | 3.6 | 2.5 | 2.1 | 3.0 | 1.3 | 2.6 | 1.1 |
| 6. MCP interoperability and workflows / 5 | 4.3 | 4.0 | 3.8 | 3.5 | 3.0 | 3.8 | 1.3 | 3.5 |
| 7. Analytical Backbone / 5 | 3.5 | 3.8 | 3.8 | 3.5 | 3.3 | 2.0 | 2.3 | 1.5 |
| 8. Enterprise-Grade Platform / 6 | 4.5 | 2.0 | 1.5 | 3.0 | 2.0 | 2.8 | 1.8 | 1.3 |
Research takeaways by vendor
- Sellforte: 38.3/48. Highest total score and the highest scores in five categories: Marketing Data Reporting via MCP, Historical Performance Insights & Causal Explanation via MCP, Campaign & Ad Set-Level Optimization with MCP, MCP interoperability and workflows and Enterprise-Grade Platform. Sellforte is best for retailers and ecommerce businesses looking for enterprise-grade MMM and incrementality testing, and granular campaign & ad set level optimization.
- Recast: 31.4/48. Second overall, with the highest scores for Channel-Level Optimization with MCP and Incrementality Testing with MCP. It also ties Triple Whale for the highest score in Analytical Backbone.
- Triple Whale: 30.6/48. Third overall and tied with Recast for the highest score in Analytical Backbone. It ranks second for Marketing Data Reporting via MCP and third for Channel-Level Optimization with MCP.
- Lifesight: 29.8/48. Fourth overall, with the second-highest scores for Campaign & Ad Set-Level Optimization with MCP and Enterprise-Grade Platform. Its highest normalized category scores are in Historical Performance Insights & Causal Explanation via MCP, followed by Channel-Level Optimization with MCP.
- Measured: 26.6/48. Fifth overall and second for Incrementality Testing with MCP. Its highest normalized category scores are in Marketing Data Reporting via MCP and Historical Performance Insights & Causal Explanation via MCP, followed by Analytical Backbone.
- Fospha: 24.3/48. Sixth overall. Its highest normalized category scores are in MCP interoperability and workflows, followed by Marketing Data Reporting via MCP and Channel-Level Optimization with MCP.
- Haus: 18.9/48. Seventh overall. Its highest normalized category scores are in Incrementality Testing with MCP and Marketing Data Reporting via MCP, followed by Analytical Backbone.
- Northbeam: 17.6/48. Eighth overall. Its highest normalized category scores are in Marketing Data Reporting via MCP and MCP interoperability and workflows.
Introduction and Table of Contents
In 2026, Marketing Mix Modeling (MMM) and incrementality testing vendors have started adding Model Context Protocol (MCP) servers to their solutions, with launches from Measured and Recast among the examples. These connections let marketers bring their measurement data into AI assistants such as Claude and ChatGPT, so they can ask questions and work with results in the tools they already use.
The capabilities available through these connections differ. A marketer might want to retrieve performance data, understand what drove sales, calculate a new budget scenario or review an incrementality test. The depth of support for each task varies by vendor, as does the level of detail available for channels, campaigns and ad sets. Buyers need to understand what each solution makes available through MCP and how well it supports their day-to-day decisions.
We conducted this evaluation to bring greater transparency to the market. Using a common framework of 48 criteria across eight categories, Claude and ChatGPT assessed public documentation for eight vendors. By publishing the criteria, scoring method and results, we aim to make the differences easier to compare and help marketers identify the capabilities they should examine more closely when choosing a solution.
- What is an MCP server for MMM and incrementality testing?
- Research methodology: evaluation criteria
- Research methodology: scoring
- Which vendors were evaluated?
- Open-source and self-built MCP servers for MMM
- In-depth comparison by category
- Summary by vendor
- Frequently asked questions
- Change log
- Evaluation dates by vendor and LLM
- Limitations and disclosures
- Further reading and resources
What Is an MCP Server for MMM and Incrementality Testing?
An MCP server for Marketing Mix Modeling (MMM) and incrementality testing enables marketers to interact with their marketing measurement solution through an AI assistant such as Claude or ChatGPT. They can ask questions in plain language and work with their company’s measurement data and supported analytical tools directly in the conversation.
For MMM, this can mean asking which channels drove incremental sales, how campaign-level incremental ROAS compares with platform-reported ROAS, or how a different budget allocation would affect expected revenue. For incrementality testing, it can mean retrieving the results of a completed geo test, reviewing measured lift or identifying which experiment to run next, depending on the capabilities the vendor makes available.
Through the MCP connection, the assistant requests data or calls a supported tool in the measurement solution. The solution returns the relevant model outputs, test results or scenario calculations, which the assistant can explain and use in further analysis. The functions available through MCP vary by vendor, which is why this evaluation examines individual capabilities.
Dedicated MCP servers for MMM and incrementality testing are available today. Measurement vendors including Sellforte, Recast, Triple Whale, Lifesight, Measured, Fospha and Northbeam publish documentation for MCP access to their platforms, and open-source MCP servers exist for teams that run their own models.
Research Methodology: Evaluation Criteria
This article follows Sellforte's Solution Research Objectives and Guiding principles: We first conducted primary research to form an evaluation framework. We then gave that framework to Claude and ChatGPT for conducting the vendor evaluation.
To develop the 48 criteria in the evaluation framework on this article, we:
- Analyzed 5,776 prompts that marketers used in interacting with Sellforte via MCP or Sellforte's in-platform conversational AI interface. These prompts gave us an understanding of the real usage patterns and expectations that marketers have towards MCPs.
- Analyzed 570 discussions with marketers, where MCP and conversational AI interfaces in MMM and incrementality testing was discussed.
- Reviewed our experience in more than 20 retail and ecommerce RFPs and tender processes that contained requirements for MCPs and AI capabilities of MMM and incrementality testing platforms.
- Used desk research and LLM-assisted investigation to identify gaps and check the framework against public information.
The framework has 48 criteria in eight categories. Each criterion is worth up to one point, so the maximum total score is 48.
- Marketing Data Reporting via MCP (6 criteria): access to actual business and media data, including online and physical-store sales, digital and offline media, and filtering by planning dimensions.
- Historical Performance Insights & Causal Explanation via MCP (5 criteria): access to modeled incremental returns, promotion effects and the drivers of performance.
- Channel-Level Optimization with MCP (8 criteria): the planning workflow from a proposed budget to an optimized allocation, forecast, constraints, scenario comparison and saved plan.
- Campaign & Ad Set-Level Optimization with MCP (8 criteria): incremental and marginal returns, budget and bid recommendations, execution through ad-platform APIs and post-change reporting.
- Incrementality Testing with MCP (5 criteria): access to geo test, own-media A/B test and Meta Conversion Lift results, plus prioritizing and designing new experiments.
- MCP interoperability and workflows (5 criteria): setup guidance, tool and metric documentation, reusable structured data and recurring workflows in external agents.
- Analytical Backbone (5 criteria): the measurement system behind the answers, including a Bayesian MMM foundation, experiment calibration and model validation.
- Enterprise-Grade Platform (6 criteria): enterprise references, independent security assurance, residency and cloud options, single sign-on and public hands-on MCP access.
The full list of 48 criteria, with a description of each, is in the MCP server evaluation checklist for MMM and incrementality testing. Criterion-level scores for every vendor are in the in-depth comparison below.
Research Methodology: Scoring
Scoring each tool against the evaluation criteria was done by Claude and ChatGPT. See specific LLM-model versions in Evaluation Dates by Vendor and LLM section. Here were the assessment steps:
Step 1. Both LLMs independently scored each criterion for each vendor, based on the instructions in this section.
Step 2. For the final score for each criteria, we used an average between the two LLMs, rounded to 1 decimal.
Step 3. Category scores were created by summing up the scores from each criteria in the category, and total score was calculated by summing up the category scores.
The full research notes include the evaluation criteria, instructions and individual Claude and ChatGPT evaluations for each vendor.
Why is the scoring done by Claude and ChatGPT?
Simulating evaluation a buyer might make based on public materials. By using only vendor-provided materials on their website and technical documentation in the assessment, Claude and ChatGPT -based evaluation simulates how a buyer without prior knowledge of the vendor might assess the vendor prior to a sales call or demo meeting.
Transparent methodology and reproducible assessment. Because the evaluation instructions are published alongside this article, any reader can re-run the evaluation for any vendor and verify or challenge the results. This is a higher standard of transparency than conventional analyst-style research, where the scoring rationale is typically not disclosed.
Equal treatment of vendors. Claude and ChatGPT apply the same instructions, the same criteria, the same scoring scale, and the same source prioritization rules to every vendor in the comparison.
Instructions given to Claude and ChatGPT
The following instructions were given to Claude and ChatGPT for conducting the evaluation.
| ID | Instruction |
|---|---|
| 1 | You are an independent evaluator of MCP servers for Marketing Mix Modeling and Incrementality testing |
| 2 | Use these evaluation criteria for categories 1-5: 1.00: Strong public evidence that the capability exists in the MCP 0.75: Partial public evidence that the capability exists in the MCP 0.50: No evidence found that the capability is available via the MCP, but there is strong evidence that the capability is available in the broader platform 0.25: No evidence found that the capability is available via the MCP, but there is partial evidence that the capability is available in the broader platform 0.00 : No evidence found that the capability exists in the MCP or in the broader platform Use these evaluation criteria for categories 6-8: 1: Strong evidence that the platform supports the capability 0.5: Partial evidence that the platform supports the capability 0: No evidence found that the platform supports the capability |
| 3 | In the evaluation, only use information available at the company website, and in the domain where technical documentation is located (if separately hosted). |
| 4 | Prioritize type of source materials in this order 1. Technical documentation (such as support center) 2. Product page 3. Marketing collateral (such as product launch blog posts) |
| 5 | When interpreting terminology, you can assume that - ROI is the same thing as incremental ROAS or iROAS - Marginal ROI is the same thing as Marginal Incremental ROAS or miROAS - Diminishing return curves are the same thing as response curves |
| 6 | For the purposes of this evaluation, feature available via API also qualify with same scores as features via MCP. |
| 7 | As an output, provide your evaluation in an excel file, using these columns: - Category ID - Category - Criteria ID - Criteria - Criteria Score - Two-sentence rationale for the score. If you found no evidence, comment that you did not find evidence, instead of claiming that the capability does not exist - URL to source - Type of source (Technical doc, Product page, Marketing collateral) |
As you can see from the scoring model, the LLMs are instructed to evaluate publicly available documentation. This means that low scores or zero scores may reflect the absence of publicly documentation, not necessarily absence of the capability itself.
Why do different categories use different scoring scales?
Categories 1–5 assess the measurement and optimization capabilities that marketers can access through the MCP. Scores 1.0 or 0.75 are used if there is strong or partial evidence that the capability is available via the MCP.
We also asked the LLMs to give a score of 0.5 or 0.25 if there is evidence for a capability being available in the overall platform, but there is no evidence for it being available via the MCP. As an example, a platform might be able to show geo test results, but there is no documentation where the MCP can summarize their findings. Why did we apply make this decision? Firstly, it is possible that the capability is already available via the MCP, but the documentation does not yet exist. Secondly, if the capability is not in MCP yet, it is likely that it will be added to the MCP in the near future.
Categories 6–8 assess MCP interoperability and workflows, Analytical Backbone, and Enterprise-Grade Platform. For these criteria, the question is whether there is evidence for the overall solution supporting the specified requirements. Thus, a simpler three-level scale was used: 0, 0.5. 1.0.
Why do API capabilities qualify for the same scores as MCP capabilities?
We included equivalent API capabilities because the evaluation aims to assess the measurement data and functions that can support AI-assisted marketing workflows. An MCP tool can call an existing API to retrieve data or run an operation, as described in the official MCP tools specification. A documented API can therefore provide a route to the same underlying capability through an integration. Giving it equivalent credit allows us to recognize that access even when the vendor has not exposed the capability through its own MCP server.
This is a scoring choice for this research. An API must provide evidence of the capability being assessed; simply offering an API is not enough. API-supported functionality may require additional integration work before a marketer can use it through an AI assistant, so an equivalent score does not imply an identical setup experience.
How to reproduce the evaluation and scoring yourself with Claude or ChatGPT?
To reproduce the evaluation for any vendor in this study, you can follow the instructions in this section.
Step 1. Open Claude or ChatGPT incognito mode. Choose the model and level of effort.
Step 2. Provide it following context:
-
Evaluation instructions that have been shared in this article
-
Evaluation criteria, listed in the research notes and described in the evaluation checklist
Step 3. Initiate the prompt: "Evaluate [add vendor name, e.g., Sellforte], based on the instructions and criteria provided as context for this prompt."
What biases and limitations does our scoring approach have, and how are we addressing them?
While LLM-based evaluation reduces vendor bias and promotes equal treatment of vendors and reproducibility, there are four main limitations to be aware of.
1. Availability of documentation that each vendor has made public. A vendor with extensive and detailed public documentation will naturally score higher than one that keeps product details behind a sales call gate, even if the underlying capabilities are comparable. Each vendor can affect their own scoring through public documentation.
2. LLMs' ability to find public documentation. Even with web search enabled, Claude and ChatGPT may not find every relevant page on a vendor's website. Documentation that is not well-indexed or that sits in obscure subdomains may be missed. We tried to address this by using advanced models from two separate LLMs, and added a specific point to the instructions to search for vendor-provided technical documentation that might sometimes be under a different sub-domain.
3. LLMs' ability to interpret public documentation against the scoring criteria. Matching a product description to a specific criterion requires judgment. We reduce interpretation variance by using advanced models from two separate LLMs, providing detailed criterion descriptions, and providing explicit terminology equivalences. However, edge cases may still exist.
4. Limitations of LLM technology, including hallucinations. LLMs are known to occasionally assert things that are not based on facts. To reduce this risk, we used advanced models from two separate LLMs, asked them to provide source URL for their assessment, and asked them to provide a rationale for the score in each criterion.
Which MCP Vendors Were Evaluated?
We focused this article on MCP servers from recognized vendors that provide Marketing Mix Modeling and/or incrementality testing. When searching for vendors that would meet these criteria we looked into multiple product categories: Data connector companies, traditional Marketing Mix Modeling providers, next gen MMM vendors, Incrementality testing tools, attribution tools.
We evaluated eight MCP servers for MMM and incrementality testing: Sellforte, Recast, Triple Whale, Lifesight, Measured, Fospha, Haus and Northbeam.
- Sellforte MCP: connects all Sellforte product modules (MMM results, incrementality tests, budget optimization and campaign recommendations) to Claude, ChatGPT, Gemini and other MCP clients.
- Recast MCP: a hosted MCP server for Recast's Bayesian MMM platform, with setup guides for ChatGPT, Claude and other clients.
- Triple Whale MCP: a read-only connection that lets AI tools such as Claude and ChatGPT query real-time Triple Whale data.
- Lifesight MCP: read-only access from Claude and ChatGPT to MMM, geo incrementality testing and attribution data in Lifesight.
- Measured MCP: brings Measured's cross-channel performance and incrementality data into ChatGPT, Claude, Gemini and other AI platforms.
- Fospha MCP: connects Fospha's measurement data to AI tools such as Claude and ChatGPT.
- Haus: describes an MCP that makes its experiment, MMM and attribution data available to LLM tools. No public MCP documentation was found.
- Northbeam MCP: a read-only connection for querying Northbeam dashboards from Claude, ChatGPT and other assistants.
Descriptions, documentation links and scores for each vendor are in the summary by vendor.
Open-Source and Self-Built MCP Servers for MMM
This research scores vendor platforms. Teams that run their own marketing mix model can also connect it to an AI assistant with an open-source or self-built MCP server. These options were not scored.
- mcp-server-meridian (published by BlueAlpha, MIT license): exposes a fitted Google Meridian model as MCP tools covering ROI, saturation, adstock, priors and budget simulation.
- simba-mcp (published by Simba, MIT license): connects AI assistants to Simba's Bayesian MMM platform to upload data, build models, check results and run budget optimizations.
- A custom MCP server: wraps the query functions of an in-house model, for example one built with Google Meridian, PyMC-Marketing or Meta's Robyn, so that the assistant calls approved functions instead of writing its own analysis code.
With these options, your team is responsible for the model, its validation and its upkeep. A vendor MCP server adds a maintained measurement platform, which is what the 48 criteria in this research assess.
In-depth Comparison of MCP Servers for MMM and Incrementality Testing
The charts and tables in this section show the research scores and the results for all research categories
For the individual Claude and ChatGPT evaluations behind these results, see the full research notes.
How to read the scores: each score is the average of the Claude and ChatGPT evaluations of a vendor's public documentation, reviewed on September 23, 2026. The scores are not product tests. A low or zero score can reflect limited public documentation and does not show that a capability is missing.
Total research score across all 48 criteria
Bars and ranking use unrounded scores; labels are rounded to one decimal.
Sellforte scored 38.3/48, followed by Recast at 31.4, Triple Whale at 30.6, Lifesight at 29.8, Measured at 26.6, Fospha at 24.3, Haus at 18.9, and Northbeam at 17.6.
The average across the eight vendors is 27.2/48. The three categories with the greatest variation in vendor scores were
-
Channel-Level Optimization with MCP (score range 2.3 - 7.9 out of 8)
-
Historical Performance Insights & Causal Explanation via MCP (score range 1.6 - 5.0 out of 5)
-
MCP interoperability and workflows (score ranged 1.3 - 4.3 out of 5)
| Evaluated category | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Max |
|---|---|---|---|---|---|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 6.0 | 5.0 | 5.3 | 4.1 | 4.9 | 4.3 | 3.1 | 5.0 | 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 5.0 | 4.1 | 3.6 | 3.8 | 3.6 | 2.4 | 1.9 | 1.6 | 5 |
| 3. Channel-Level Optimization with MCP | 7.4 | 7.9 | 6.5 | 5.6 | 4.5 | 4.6 | 2.9 | 2.3 | 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 5.0 | 1.0 | 3.8 | 4.1 | 2.4 | 3.3 | 3.1 | 1.4 | 8 |
| 5. Incrementality Testing with MCP | 2.6 | 3.6 | 2.5 | 2.1 | 3.0 | 1.3 | 2.6 | 1.1 | 5 |
| 6. MCP interoperability and workflows | 4.3 | 4.0 | 3.8 | 3.5 | 3.0 | 3.8 | 1.3 | 3.5 | 5 |
| 7. Analytical Backbone | 3.5 | 3.8 | 3.8 | 3.5 | 3.3 | 2.0 | 2.3 | 1.5 | 5 |
| 8. Enterprise-Grade Platform | 4.5 | 2.0 | 1.5 | 3.0 | 2.0 | 2.8 | 1.8 | 1.3 | 6 |
| Total | 38.3 | 31.4 | 30.6 | 29.8 | 26.6 | 24.3 | 18.9 | 17.6 | 48 |
1. Marketing Data Reporting via MCP: which vendor scored highest?
Highest research scores for Marketing Data Reporting via MCP: Sellforte (6.0 out of 6), followed by Triple Whale (5.3), then Recast and Northbeam (5.0 each).
What we evaluated. The six criteria assess reporting through MCP: online and offline store sales, digital and offline media data, business outcomes beyond revenue, and filtering or grouping results by business dimensions.
What the research found. The research scores were: Sellforte (6.0/6), Triple Whale (5.3/6), Recast (5.0/6), Northbeam (5.0/6), Measured (4.9/6), Fospha (4.3/6), Lifesight (4.1/6), Haus (3.1/6). Recast and Northbeam are tied.
Highest average research scores were for reporting business outcomes beyond revenue and digital media data. Lower average research scores were for offline store sales and offline media reporting. These two reporting criteria also showed the greatest variation among vendors, with offline store sales varying most.
| Criterion | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Average research score |
|---|---|---|---|---|---|---|---|---|---|
| 1.1 MCP reports sales progress for online sales | 1.0 | 0.9 | 1.0 | 0.6 | 0.8 | 1.0 | 0.6 | 1.0 | 0.9 |
| 1.2 MCP reports sales progress for offline store sales | 1.0 | 0.9 | 0.8 | 0.5 | 0.8 | 0.1 | 0.4 | 0.5 | 0.6 |
| 1.3 MCP reports digital media data (spend, media metrics) | 1.0 | 0.8 | 1.0 | 0.9 | 1.0 | 1.0 | 0.8 | 1.0 | 0.9 |
| 1.4 MCP reports offline media data (spend, media metrics) | 1.0 | 0.8 | 0.6 | 0.5 | 0.8 | 0.3 | 0.4 | 0.8 | 0.6 |
| 1.5 MCP Reports business outcomes beyond revenue | 1.0 | 1.0 | 1.0 | 0.9 | 0.9 | 1.0 | 0.8 | 1.0 | 0.9 |
| 1.6 Filters and groups by business dimensions | 1.0 | 0.8 | 0.9 | 0.8 | 0.8 | 0.9 | 0.3 | 0.8 | 0.8 |
| Category total | 6.0 | 5.0 | 5.3 | 4.1 | 4.9 | 4.3 | 3.1 | 5.0 | 4.7 |
2. Historical Performance Insights & Causal Explanation via MCP: which vendor scored highest?
Highest research scores for Historical Performance Insights & Causal Explanation via MCP: Sellforte (5.0 out of 5), followed by Recast (4.1), then Lifesight (3.8).
What we evaluated. The five criteria assess historical performance insights through MCP, including incremental ROAS and revenue for digital and offline channels, promotion-driven revenue, daily MMM-based measurement updates, and the decomposition of base, media and non-media performance drivers.
What the research found. The research scores were: Sellforte (5.0/5), Recast (4.1/5), Lifesight (3.8/5), Triple Whale (3.6/5), Measured (3.6/5), Fospha (2.4/5), Haus (1.9/5), Northbeam (1.6/5). Triple Whale and Measured are tied.
Highest average research scores were for reporting incremental ROAS and revenue for digital channels and decomposing performance drivers. Lower average research scores were for daily MMM-based updates to incremental ROAS. Promotion-driven revenue reporting and daily MMM-based updates showed the greatest variation among vendors.
| Criterion | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Average research score |
|---|---|---|---|---|---|---|---|---|---|
| 2.1 MCP reports incremental ROAS and incremental revenue for each digital channel | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.8 | 0.5 | 0.9 |
| 2.2 MCP reports incremental ROAS and incremental revenue for each offline channel | 1.0 | 1.0 | 0.6 | 0.6 | 0.8 | 0.1 | 0.5 | 0.3 | 0.6 |
| 2.3 MCP reports promotion-driven revenue, in addition to media-driven | 1.0 | 1.0 | 1.0 | 0.8 | 0.5 | 0.1 | 0.3 | 0.4 | 0.6 |
| 2.4 MCP's incremental ROAS measurement is updated daily (not weekly or monthly), based on MMM | 1.0 | 0.1 | 0.1 | 0.5 | 0.6 | 0.9 | 0.1 | 0.3 | 0.5 |
| 2.5 Decomposes performance drivers (Base, media, non-media drivers) | 1.0 | 1.0 | 0.9 | 0.9 | 0.8 | 0.3 | 0.3 | 0.3 | 0.7 |
| Category total | 5.0 | 4.1 | 3.6 | 3.8 | 3.6 | 2.4 | 1.9 | 1.6 | 3.3 |
3. Channel-Level Optimization with MCP: which vendor scored highest?
Highest research scores for Channel-Level Optimization with MCP: Recast (7.9 out of 8), followed by Sellforte (7.4), then Triple Whale (6.5).
What we evaluated. The eight criteria assess channel-level planning through MCP: creating optimized allocations, forecasting media plans, reporting marginal returns and response curves, simulating budget changes, applying planning constraints, comparing scenarios, planning against business targets, and retrieving saved plans and assumptions.
What the research found. The research scores were: Recast (7.9/8), Sellforte (7.4/8), Triple Whale (6.5/8), Lifesight (5.6/8), Fospha (4.6/8), Measured (4.5/8), Haus (2.9/8), Northbeam (2.3/8).
Highest average research scores were for simulating a user-specified budget change and forecasting a media plan. Lower average research scores were for enforcing planning constraints and retrieving saved plans and assumptions, which tied for the lowest average. Those two criteria also showed the greatest variation among vendors, with retrieval of saved plans and assumptions varying most.
| Criterion | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Average research score |
|---|---|---|---|---|---|---|---|---|---|
| 3.1 MCP creates an optimized channel allocation | 1.0 | 1.0 | 0.8 | 0.8 | 0.5 | 0.5 | 0.5 | 0.5 | 0.7 |
| 3.2 MCP forecasts a specified media plan | 1.0 | 1.0 | 0.8 | 0.8 | 0.5 | 0.8 | 0.5 | 0.5 | 0.7 |
| 3.3 MCP reports marginal returns and response curves | 0.5 | 1.0 | 0.8 | 0.8 | 0.8 | 0.9 | 0.6 | 0.3 | 0.7 |
| 3.4 MCP simulates a user-specified budget change | 1.0 | 1.0 | 0.8 | 0.8 | 0.5 | 1.0 | 0.5 | 0.5 | 0.8 |
| 3.5 MCP enforces planning constraints | 1.0 | 1.0 | 0.8 | 0.5 | 0.5 | 0.5 | 0.3 | 0.0 | 0.6 |
| 3.6 MCP compares scenarios with a baseline plan | 1.0 | 0.9 | 1.0 | 0.8 | 0.6 | 0.4 | 0.3 | 0.3 | 0.6 |
| 3.7 MCP plans against a business target | 0.9 | 1.0 | 0.8 | 0.8 | 0.5 | 0.4 | 0.3 | 0.3 | 0.6 |
| 3.8 MCP retrieves saved plans and assumptions | 1.0 | 1.0 | 1.0 | 0.6 | 0.6 | 0.3 | 0.0 | 0.0 | 0.6 |
| Category total | 7.4 | 7.9 | 6.5 | 5.6 | 4.5 | 4.6 | 2.9 | 2.3 | 5.2 |
4. Campaign & Ad Set-Level Optimization with MCP: which vendor scored highest?
Highest research scores for Campaign & Ad Set-Level Optimization with MCP: Sellforte (5.0 out of 8), followed by Lifesight (4.1), then Triple Whale (3.8).
What we evaluated. The eight criteria assess campaign and ad set optimization through MCP, from reporting incremental ROAS and comparing it with attribution-based ROAS to providing marginal returns, recommending daily budgets and bid values, pushing budget and bidding changes to ad platforms, and analyzing the revenue and spend impact of executed bidding changes.
What the research found. The research scores were: Sellforte (5.0/8), Lifesight (4.1/8), Triple Whale (3.8/8), Fospha (3.3/8), Haus (3.1/8), Measured (2.4/8), Northbeam (1.4/8), Recast (1.0/8).
Highest average research scores were for comparing incremental ROAS with last-click and ad platform attribution ROAS, and reporting incremental ROAS by campaign and ad set. Lower average research scores were for recommending optimal bid values and pushing bidding changes to ad platforms, which tied for the lowest average. Campaign and ad set incremental ROAS reporting and its comparison with attribution-based ROAS showed the greatest variation among vendors.
| Criterion | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Average research score |
|---|---|---|---|---|---|---|---|---|---|
| 4.1 MCP reports incremental ROAS for each campaign and ad set | 1.0 | 0.3 | 0.8 | 0.9 | 0.8 | 0.8 | 0.8 | 0.3 | 0.7 |
| 4.2 For each campaign and ad set, MCP compares incremental ROAS to last-click and ad platform attribution ROAS | 1.0 | 0.1 | 0.8 | 0.8 | 0.8 | 0.9 | 0.5 | 0.8 | 0.7 |
| 4.3 For each campaign and ad set, MCP provides marginal returns (miROAS) | 0.5 | 0.0 | 0.5 | 0.5 | 0.4 | 0.1 | 0.3 | 0.0 | 0.3 |
| 4.4 For each campaign and ad set, MCP recommends optimal daily spend/budget | 0.5 | 0.5 | 0.4 | 0.5 | 0.3 | 0.4 | 0.4 | 0.1 | 0.4 |
| 4.5 For each campaign and ad set, MCP recommends optimal bid value (e.g., Target ROAS) | 0.5 | 0.0 | 0.1 | 0.3 | 0.0 | 0.1 | 0.3 | 0.3 | 0.2 |
| 4.6 MCP can push daily spend/budget changes to Meta, Google etc. APIs | 0.5 | 0.0 | 0.5 | 0.5 | 0.1 | 0.4 | 0.6 | 0.0 | 0.3 |
| 4.7 MCP can push bidding changes (e.g., Target ROAS) to Meta, Google etc. APIs | 0.5 | 0.0 | 0.3 | 0.5 | 0.0 | 0.1 | 0.1 | 0.0 | 0.2 |
| 4.8 For each executed bidding change for a campaign or ad set, MCP provides pre/post analysis summarizing revenue and spend impact of the change | 0.5 | 0.1 | 0.5 | 0.3 | 0.1 | 0.5 | 0.3 | 0.0 | 0.3 |
| Category total | 5.0 | 1.0 | 3.8 | 4.1 | 2.4 | 3.3 | 3.1 | 1.4 | 3.0 |
5. Incrementality Testing with MCP: which vendor scored highest?
Highest research scores for Incrementality Testing with MCP: Recast (3.6 out of 5), followed by Measured (3.0), then Sellforte and Haus (2.6 each).
What we evaluated. The five criteria assess incrementality testing through MCP: reporting results from geo tests, own-media A/B tests and Meta Conversion Lift tests, prioritizing experiments, and designing feasible incrementality tests.
What the research found. The research scores were: Recast (3.6/5), Measured (3.0/5), Sellforte (2.6/5), Haus (2.6/5), Triple Whale (2.5/5), Lifesight (2.1/5), Fospha (1.3/5), Northbeam (1.1/5). Sellforte and Haus are tied.
Highest average research scores were for reporting geo-test results and prioritizing experiments. Lower average research scores were for reporting own-media A/B test results and Meta Conversion Lift findings, which tied for the lowest average. Those two reporting criteria also showed the greatest variation among vendors, with own-media A/B test reporting varying most.
| Criterion | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Average research score |
|---|---|---|---|---|---|---|---|---|---|
| 5.1 MCP reports results for Geo Tests | 0.8 | 0.9 | 1.0 | 0.8 | 0.9 | 0.5 | 0.8 | 0.5 | 0.8 |
| 5.2 MCP reports results for Own Media A/B tests (e.g. leaflet tests) | 0.8 | 0.9 | 0.0 | 0.3 | 0.5 | 0.0 | 0.6 | 0.0 | 0.4 |
| 5.3 MCP reports findings for Meta Conversion Lift tests | 0.8 | 0.9 | 0.6 | 0.1 | 0.1 | 0.3 | 0.3 | 0.0 | 0.4 |
| 5.4 MCP prioritizes experiments | 0.1 | 0.5 | 0.4 | 0.5 | 1.0 | 0.3 | 0.5 | 0.3 | 0.4 |
| 5.5 MCP designs feasible incrementality tests | 0.3 | 0.5 | 0.5 | 0.5 | 0.5 | 0.3 | 0.5 | 0.4 | 0.4 |
| Category total | 2.6 | 3.6 | 2.5 | 2.1 | 3.0 | 1.3 | 2.6 | 1.1 | 2.4 |
6. MCP interoperability and workflows: which vendor scored highest?
Highest research scores for MCP interoperability and workflows: Sellforte (4.3 out of 5), followed by Recast (4.0), then Triple Whale and Fospha (3.8 each).
What we evaluated. The five criteria assess MCP interoperability and workflows: documented instructions for connecting with Claude and ChatGPT, native visual outputs, accessible tool and metric documentation, reusable structured data, and automation through external agents.
What the research found. The research scores were: Sellforte (4.3/5), Recast (4.0/5), Triple Whale (3.8/5), Fospha (3.8/5), Lifesight (3.5/5), Northbeam (3.5/5), Measured (3.0/5), Haus (1.3/5). Triple Whale and Fospha are tied, as are Lifesight and Northbeam.
Highest average research scores were for returning reusable structured data and providing connection instructions for Claude and ChatGPT. Lower average research scores were for native visual outputs. Native visual outputs and external agent automation showed the greatest variation among vendors.
| Criterion | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Average research score |
|---|---|---|---|---|---|---|---|---|---|
| 6.1 Documented instructions for connecting with Claude and ChatGPT | 0.8 | 1.0 | 1.0 | 1.0 | 0.8 | 1.0 | 0.3 | 1.0 | 0.8 |
| 6.2 MCP provides native visual outputs | 1.0 | 0.0 | 0.0 | 0.5 | 0.5 | 0.0 | 0.0 | 0.0 | 0.3 |
| 6.3 Exposes tool and metric documentation | 1.0 | 1.0 | 0.8 | 0.8 | 0.3 | 1.0 | 0.5 | 0.8 | 0.8 |
| 6.4 Returns reusable structured data | 1.0 | 1.0 | 1.0 | 0.8 | 1.0 | 1.0 | 0.5 | 1.0 | 0.9 |
| 6.5 Supports external agent automation | 0.5 | 1.0 | 1.0 | 0.5 | 0.5 | 0.8 | 0.0 | 0.8 | 0.6 |
| Category total | 4.3 | 4.0 | 3.8 | 3.5 | 3.0 | 3.8 | 1.3 | 3.5 | 3.4 |
7. Analytical Backbone: which vendor scored highest?
Highest research scores for Analytical Backbone: Recast and Triple Whale (3.8 out of 5 each), followed by Sellforte and Lifesight (3.5 each).
What we evaluated. The five criteria assess the analytical foundations behind MCP responses: deterministic answers backed by Bayesian MMM, calibration with incrementality tests, reporting model validation and modelling KPIs, auditable and editable calibration settings in a self-serve interface, and controlled model calibration through MCP.
What the research found. The research scores were: Recast (3.8/5), Triple Whale (3.8/5), Sellforte (3.5/5), Lifesight (3.5/5), Measured (3.3/5), Haus (2.3/5), Fospha (2.0/5), Northbeam (1.5/5). Recast and Triple Whale are tied, as are Sellforte and Lifesight.
Highest average research scores were for calibrating the underlying MMM with incrementality tests and reporting model validation and modelling KPIs. Lower average research scores were for controlled calibration through MCP. Model validation reporting and auditable, editable calibration settings in a self-serve interface showed the greatest variation among vendors.
| Criterion | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Average research score |
|---|---|---|---|---|---|---|---|---|---|
| 7.1 MCP provides deterministic, model-backed answers with Bayesian MMM as backbone | 1.0 | 1.0 | 0.8 | 0.5 | 0.5 | 0.5 | 0.5 | 0.3 | 0.6 |
| 7.2 The underlying MMM used by the MCP is calibrated with incrementality tests | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.5 | 1.0 | 0.8 | 0.9 |
| 7.3 MCP reports model validation and other modelling KPIs | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.8 | 0.5 | 0.0 | 0.8 |
| 7.4 Model calibration & configuration settings (e.g., priors) are auditable and editable in a self-serve UI | 0.5 | 0.5 | 1.0 | 1.0 | 0.8 | 0.3 | 0.3 | 0.5 | 0.6 |
| 7.5 Supports controlled calibration through MCP | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Category total | 3.5 | 3.8 | 3.8 | 3.5 | 3.3 | 2.0 | 2.3 | 1.5 | 2.9 |
8. Enterprise-Grade Platform: which vendor scored highest?
Highest research scores for Enterprise-Grade Platform: Sellforte (4.5 out of 6), followed by Lifesight (3.0), then Fospha (2.8).
What we evaluated. The six criteria assess enterprise readiness: at least ten public reference customers with $1 billion or more in revenue, independent security assurance, US and EU data residency options, a choice of AWS, GCP and Azure, enterprise single sign-on, and a hands-on MCP demo or trial available without a sales call.
What the research found. The research scores were: Sellforte (4.5/6), Lifesight (3.0/6), Fospha (2.8/6), Recast (2.0/6), Measured (2.0/6), Haus (1.8/6), Triple Whale (1.5/6), Northbeam (1.3/6). Recast and Measured are tied.
Highest average research scores were for independent security assurance and enterprise single sign-on. Lower average research scores were for access to a hands-on MCP demo or trial without a sales call and a choice of cloud providers. US and EU data residency options and the choice of cloud providers showed the greatest variation among vendors.
| Criterion | Sellforte | Recast | Triple Whale | Lifesight | Measured | Fospha | Haus | Northbeam | Average research score |
|---|---|---|---|---|---|---|---|---|---|
| 8.1 At least 10 public reference customers from $1B+ revenue brands | 0.8 | 0.5 | 0.0 | 0.3 | 0.5 | 0.5 | 0.8 | 0.0 | 0.4 |
| 8.2 SOC 2, ISO 27001, or audited IT security by a third-party cyber security auditor | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
| 8.3 Data residency: geography option between US and EU | 0.8 | 0.0 | 0.0 | 1.0 | 0.0 | 0.5 | 0.0 | 0.0 | 0.3 |
| 8.4 Multi-cloud: option between AWS, GCP, and Azure | 1.0 | 0.0 | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.2 |
| 8.5 Supports single sign-on (SSO) for enterprises | 1.0 | 0.5 | 0.5 | 0.5 | 0.5 | 0.8 | 0.0 | 0.3 | 0.5 |
| 8.6 Hands-on demo or trial of the MCP capabilities is available without sales-call gating | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Category total | 4.5 | 2.0 | 1.5 | 3.0 | 2.0 | 2.8 | 1.8 | 1.3 | 2.3 |
What the Research Suggests About the Market
Bars and rankings use unrounded scores; labels are rounded to one decimal.
The three strongest areas on average were Marketing Data Reporting via MCP (78.4% of available points), MCP interoperability and workflows (67.5%), and Channel-Level Optimization with MCP (65.0%). These are table stakes capabilities that marketers should expect from a vendor.
The areas with lowest research scores on average were Campaign & Ad Set-Level Optimization with MCP (37.5% of available points), Enterprise-Grade Platform (39.1%), and Incrementality Testing with MCP (47.2%). These are the capabilities where vendors have the most opportunity to improve their offering and differentiate themselves from other vendors.
Summary by Vendor
Each profile starts with a short description of the vendor's MCP, based on the vendor's public documentation as checked on November 15, 2026, followed by its research scores from the September 23, 2026 evaluation.
For the individual Claude and ChatGPT evaluations behind these results, see the full research notes.
1. Sellforte MCP (research score: 38.3 out of 48)
What is Sellforte MCP?
Sellforte MCP is Sellforte's MCP server for Marketing Mix Modeling (MMM) and incrementality testing. It connects all Sellforte product modules to LLMs and AI agents that support MCP, so marketers can analyze measurement results, work with incrementality tests, run budget scenarios with the optimizer and act on recommendations from their AI assistant. Sellforte lists Claude, ChatGPT and Gemini as supported assistants.
For incrementality testing, Sellforte's documentation describes access to Conversion Lift, GeoLift and A/B test results, including comparing experiments and viewing them alongside MMM results for the same channel. That documentation, Incrementality testing with Sellforte MCP, was published on October 1, 2026, after the evaluation date, so it is not reflected in the scores.
Best for, based on the scores in this research: retailers and ecommerce businesses looking for enterprise-grade MMM and incrementality testing, and granular campaign & ad set level optimization. Sellforte received the highest scores for campaign and ad set-level optimization and enterprise-grade platform.
Sellforte received an average research score of 38.3 out of 48, the highest total in this eight-vendor comparison.
Public documentation cited in this research: Sellforte MCP overview, Marketing reporting with Sellforte MCP, Budget optimization and scenario planning with Sellforte MCP, Connecting Sellforte MCP to your LLM client. Criterion-level evidence for Sellforte is in the full research notes.
Sellforte: research score by category
| Category | Sellforte | Average research score across vendors | Highest research score among the vendors |
|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 6.0 / 6 | 4.7 / 6 | 6.0 / 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 5.0 / 5 | 3.3 / 5 | 5.0 / 5 |
| 3. Channel-Level Optimization with MCP | 7.4 / 8 | 5.2 / 8 | 7.9 / 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 5.0 / 8 | 3.0 / 8 | 5.0 / 8 |
| 5. Incrementality Testing with MCP | 2.6 / 5 | 2.4 / 5 | 3.6 / 5 |
| 6. MCP interoperability and workflows | 4.3 / 5 | 3.4 / 5 | 4.3 / 5 |
| 7. Analytical Backbone | 3.5 / 5 | 2.9 / 5 | 3.8 / 5 |
| 8. Enterprise-Grade Platform | 4.5 / 6 | 2.3 / 6 | 4.5 / 6 |
| Total score out of 48 | 38.3 / 48 | 27.2 / 48 | 38.3 / 48 |
Percentage of available points
Bars show the score as a percentage of the category maximum.
Sellforte: highest-scoring categories
- Marketing Data Reporting via MCP: 6.0 / 6. This was above the eight-vendor average of 4.7 / 6. Within this category, its highest-scoring criteria included reporting online sales progress, reporting offline store sales progress, and reporting digital media data.
- Historical Performance Insights & Causal Explanation via MCP: 5.0 / 5. This was above the eight-vendor average of 3.3 / 5. Within this category, its highest-scoring criteria included reporting incremental ROAS and revenue for digital channels, reporting incremental ROAS and revenue for offline channels, and reporting promotion-driven revenue.
- Channel-Level Optimization with MCP: 7.4 / 8. This was above the eight-vendor average of 5.2 / 8. Within this category, its highest-scoring criteria included creating an optimized channel allocation, forecasting a media plan, and simulating a user-specified budget change.
Sellforte: lowest-scoring categories
- Incrementality Testing with MCP: 2.6 / 5. Despite being one of Sellforte’s relatively lower-scoring categories, its score was still above the eight-vendor average of 2.4 / 5.
- Campaign & Ad Set-Level Optimization with MCP: 5.0 / 8. Despite being one of Sellforte’s relatively lower-scoring categories, its score was still above the eight-vendor average of 3.0 / 8.
- Analytical Backbone: 3.5 / 5. Despite being one of Sellforte’s relatively lower-scoring categories, its score was still above the eight-vendor average of 2.9 / 5.
Notable Sellforte Reference Customers
Sellforte lists following companies as examples of public reference customers:
- Fashion Ecommerce: bonprix, Azzas 2154, Represent, Odlo
- Home & Furniture Ecommerce: Finnish Design Shop
- Specialty Ecommerce: FCP Euro, Smartphoto
- Grocery Retail: Lidl
- Fashion Retail: C&A, KIK
- Cosmetics Retail: Douglas
- Sport Retail: Interpsort
- Pet Retail: Fressnapf, Musti Group
- Specialty Retail: Tchibo
- Electronics Retail: Verkkokauppa.com
- Other segments: Telenor (Telecommunications), Paysafe (Payments), eBilet (part of Allegro Group, Events)
Sellforte pricing
Sellforte MCP is included in all Sellforte plans by default. Sellforte pricing starts at $1,900 per month for the Incrementality Testing plan, $2,500 per month for the Incremental attribution plan, $4,500 per month for Full funnel MMM plan, and $8,500 for the Enterprise plan.
2. Recast MCP (research score: 31.4 out of 48)

What is Recast MCP?
Recast MCP is a hosted MCP server that connects Recast's Bayesian MMM platform to AI assistants. Recast's setup guide covers ChatGPT, Claude, Codex CLI and Snowflake Cortex, with OAuth sign-in using Recast credentials, and its companion skill files cover Recast's Forecaster, Optimizer, Plans and Reporter APIs.
Best for, based on the scores in this research: teams that prioritize channel-level budget planning and working with incrementality test results in an AI assistant. Recast received the highest scores in both of those categories.
Recast received an average research score of 31.4 out of 48, the second-highest total in this eight-vendor comparison.
Public documentation cited in this research: The Recast MCP, MCP setup guide, The Insights API. Criterion-level evidence for Recast is in the full research notes.
Recast: research score by category
| Category | Recast | Average research score across vendors | Highest research score among the vendors |
|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 5.0 / 6 | 4.7 / 6 | 6.0 / 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 4.1 / 5 | 3.3 / 5 | 5.0 / 5 |
| 3. Channel-Level Optimization with MCP | 7.9 / 8 | 5.2 / 8 | 7.9 / 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 1.0 / 8 | 3.0 / 8 | 5.0 / 8 |
| 5. Incrementality Testing with MCP | 3.6 / 5 | 2.4 / 5 | 3.6 / 5 |
| 6. MCP interoperability and workflows | 4.0 / 5 | 3.4 / 5 | 4.3 / 5 |
| 7. Analytical Backbone | 3.8 / 5 | 2.9 / 5 | 3.8 / 5 |
| 8. Enterprise-Grade Platform | 2.0 / 6 | 2.3 / 6 | 4.5 / 6 |
| Total score out of 48 | 31.4 / 48 | 27.2 / 48 | 38.3 / 48 |
Percentage of available points
Bars show the score as a percentage of the category maximum.
Recast: highest-scoring categories
- Channel-Level Optimization with MCP: 7.9 / 8. This was above the eight-vendor average of 5.2 / 8. Within this category, its highest-scoring criteria included creating an optimized channel allocation, forecasting a media plan, and reporting marginal returns and response curves.
- Marketing Data Reporting via MCP: 5.0 / 6. This was above the eight-vendor average of 4.7 / 6. Within this category, its highest-scoring criteria included reporting business outcomes beyond revenue, reporting online sales progress, and reporting offline store sales progress.
- Historical Performance Insights & Causal Explanation via MCP: 4.1 / 5. This was above the eight-vendor average of 3.3 / 5. Within this category, its highest-scoring criteria included reporting incremental ROAS and revenue for digital channels, reporting incremental ROAS and revenue for offline channels, and reporting promotion-driven revenue.
Recast: lowest-scoring categories
- Campaign & Ad Set-Level Optimization with MCP: 1.0 / 8. This was below the eight-vendor average of 3.0 / 8.
- Enterprise-Grade Platform: 2.0 / 6. This was below the eight-vendor average of 2.3 / 6.
- Incrementality Testing with MCP: 3.6 / 5. Despite being one of Recast’s relatively lower-scoring categories, its score was still above the eight-vendor average of 2.4 / 5.
3. Triple Whale MCP (research score: 30.6 out of 48)

What is Triple Whale MCP?
Triple Whale MCP is a read-only connection that lets external AI tools such as Claude, ChatGPT, Cursor and n8n query and analyze a business's real-time Triple Whale data. It connects through OAuth2 or an MCP Read API key.
Best for, based on the scores in this research: ecommerce brands that want broad marketing data reporting in their AI assistant. Triple Whale received the second-highest score for marketing data reporting and tied with Recast for the highest Analytical Backbone score.
Triple Whale received an average research score of 30.6 out of 48, the third-highest total in this eight-vendor comparison.
Public documentation cited in this research: Triple Whale MCP. Criterion-level evidence for Triple Whale is in the full research notes.
Triple Whale: research score by category
| Category | Triple Whale | Average research score across vendors | Highest research score among the vendors |
|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 5.3 / 6 | 4.7 / 6 | 6.0 / 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 3.6 / 5 | 3.3 / 5 | 5.0 / 5 |
| 3. Channel-Level Optimization with MCP | 6.5 / 8 | 5.2 / 8 | 7.9 / 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 3.8 / 8 | 3.0 / 8 | 5.0 / 8 |
| 5. Incrementality Testing with MCP | 2.5 / 5 | 2.4 / 5 | 3.6 / 5 |
| 6. MCP interoperability and workflows | 3.8 / 5 | 3.4 / 5 | 4.3 / 5 |
| 7. Analytical Backbone | 3.8 / 5 | 2.9 / 5 | 3.8 / 5 |
| 8. Enterprise-Grade Platform | 1.5 / 6 | 2.3 / 6 | 4.5 / 6 |
| Total score out of 48 | 30.6 / 48 | 27.2 / 48 | 38.3 / 48 |
Percentage of available points
Bars show the score as a percentage of the category maximum.
Triple Whale: highest-scoring categories
- Marketing Data Reporting via MCP: 5.3 / 6. This was above the eight-vendor average of 4.7 / 6. Within this category, its highest-scoring criteria included reporting online sales progress, reporting digital media data, and reporting business outcomes beyond revenue.
- Channel-Level Optimization with MCP: 6.5 / 8. This was above the eight-vendor average of 5.2 / 8. Within this category, its highest-scoring criteria included comparing scenarios with a baseline plan, retrieving saved plans and assumptions, and creating an optimized channel allocation.
- MCP interoperability and workflows: 3.8 / 5. This was above the eight-vendor average of 3.4 / 5. Within this category, its highest-scoring criteria included documenting connections with Claude and ChatGPT, returning reusable structured data, and supporting external agent automation. Analytical Backbone had the same normalized category score.
Triple Whale: lowest-scoring categories
- Enterprise-Grade Platform: 1.5 / 6. This was below the eight-vendor average of 2.3 / 6.
- Campaign & Ad Set-Level Optimization with MCP: 3.8 / 8. Despite being one of Triple Whale’s relatively lower-scoring categories, its score was still above the eight-vendor average of 3.0 / 8.
- Incrementality Testing with MCP: 2.5 / 5. Despite being one of Triple Whale’s relatively lower-scoring categories, its score was still above the eight-vendor average of 2.4 / 5.
4. Lifesight MCP (research score: 29.8 out of 48)

What is Lifesight MCP?
Lifesight MCP gives Claude and ChatGPT read-only access to a customer's live measurement data across MMM, geo incrementality testing and incrementality-adjusted attribution. Lifesight announced it in June 2026 with a library of prebuilt skills, and made it available to all Lifesight customers.
Best for, based on the scores in this research: teams that need campaign and ad set-level optimization and enterprise platform requirements. Lifesight received the second-highest scores in both of those categories.
Lifesight received an average research score of 29.8 out of 48, the fourth-highest total in this eight-vendor comparison.
Public documentation cited in this research: Lifesight MCP. Criterion-level evidence for Lifesight is in the full research notes.
Lifesight: research score by category
| Category | Lifesight | Average research score across vendors | Highest research score among the vendors |
|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 4.1 / 6 | 4.7 / 6 | 6.0 / 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 3.8 / 5 | 3.3 / 5 | 5.0 / 5 |
| 3. Channel-Level Optimization with MCP | 5.6 / 8 | 5.2 / 8 | 7.9 / 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 4.1 / 8 | 3.0 / 8 | 5.0 / 8 |
| 5. Incrementality Testing with MCP | 2.1 / 5 | 2.4 / 5 | 3.6 / 5 |
| 6. MCP interoperability and workflows | 3.5 / 5 | 3.4 / 5 | 4.3 / 5 |
| 7. Analytical Backbone | 3.5 / 5 | 2.9 / 5 | 3.8 / 5 |
| 8. Enterprise-Grade Platform | 3.0 / 6 | 2.3 / 6 | 4.5 / 6 |
| Total score out of 48 | 29.8 / 48 | 27.2 / 48 | 38.3 / 48 |
Percentage of available points
Bars show the score as a percentage of the category maximum.
Lifesight: highest-scoring categories
- Historical Performance Insights & Causal Explanation via MCP: 3.8 / 5. This was above the eight-vendor average of 3.3 / 5. Within this category, its highest-scoring criteria included reporting incremental ROAS and revenue for digital channels, decomposing base, media and non-media performance drivers, and reporting promotion-driven revenue.
- Channel-Level Optimization with MCP: 5.6 / 8. This was above the eight-vendor average of 5.2 / 8. Within this category, its highest-scoring criteria included creating an optimized channel allocation, forecasting a media plan, and reporting marginal returns and response curves.
- MCP interoperability and workflows: 3.5 / 5. This was above the eight-vendor average of 3.4 / 5. Within this category, its highest-scoring criteria included documenting connections with Claude and ChatGPT, exposing tool and metric documentation, and returning reusable structured data. Analytical Backbone had the same normalized category score.
Lifesight: lowest-scoring categories
- Incrementality Testing with MCP: 2.1 / 5. This was below the eight-vendor average of 2.4 / 5.
- Enterprise-Grade Platform: 3.0 / 6. Despite being one of Lifesight’s relatively lower-scoring categories, its score was still above the eight-vendor average of 2.3 / 6.
- Campaign & Ad Set-Level Optimization with MCP: 4.1 / 8. Despite being one of Lifesight’s relatively lower-scoring categories, its score was still above the eight-vendor average of 3.0 / 8.
5. Measured MCP (research score: 26.6 out of 48)

What is Measured MCP?
Measured launched its MCP server in May 2026. It brings Measured's cross-channel performance data, executive summaries and on-demand insights into ChatGPT, Claude, Gemini and other AI platforms. The launch release names three components: an Incrementality Intelligence Database, Omnichannel Measurement and Optimization Insights.
Best for, based on the scores in this research: teams that want to prioritize incrementality experiments through an AI assistant. Measured received the second-highest score for incrementality testing, including full credit for prioritizing experiments.
Measured received an average research score of 26.6 out of 48, the fifth-highest total in this eight-vendor comparison.
Public documentation cited in this research: Measured launches MCP server (press release), Measured MCP prompt library. Criterion-level evidence for Measured is in the full research notes.
Measured: research score by category
| Category | Measured | Average research score across vendors | Highest research score among the vendors |
|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 4.9 / 6 | 4.7 / 6 | 6.0 / 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 3.6 / 5 | 3.3 / 5 | 5.0 / 5 |
| 3. Channel-Level Optimization with MCP | 4.5 / 8 | 5.2 / 8 | 7.9 / 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 2.4 / 8 | 3.0 / 8 | 5.0 / 8 |
| 5. Incrementality Testing with MCP | 3.0 / 5 | 2.4 / 5 | 3.6 / 5 |
| 6. MCP interoperability and workflows | 3.0 / 5 | 3.4 / 5 | 4.3 / 5 |
| 7. Analytical Backbone | 3.3 / 5 | 2.9 / 5 | 3.8 / 5 |
| 8. Enterprise-Grade Platform | 2.0 / 6 | 2.3 / 6 | 4.5 / 6 |
| Total score out of 48 | 26.6 / 48 | 27.2 / 48 | 38.3 / 48 |
Percentage of available points
Bars show the score as a percentage of the category maximum.
Measured: highest-scoring categories
- Marketing Data Reporting via MCP: 4.9 / 6. This was above the eight-vendor average of 4.7 / 6. Within this category, its highest-scoring criteria included reporting digital media data, reporting business outcomes beyond revenue, and reporting online sales progress.
- Historical Performance Insights & Causal Explanation via MCP: 3.6 / 5. This was above the eight-vendor average of 3.3 / 5. Within this category, its highest-scoring criteria included reporting incremental ROAS and revenue for digital channels, reporting incremental ROAS and revenue for offline channels, and decomposing base, media and non-media performance drivers.
- Analytical Backbone: 3.3 / 5. This was above the eight-vendor average of 2.9 / 5. Within this category, its highest-scoring criteria included calibrating the underlying MMM with incrementality tests, reporting model validation and modelling KPIs, and providing auditable, editable calibration settings in a self-serve interface.
Measured: lowest-scoring categories
- Campaign & Ad Set-Level Optimization with MCP: 2.4 / 8. This was below the eight-vendor average of 3.0 / 8.
- Enterprise-Grade Platform: 2.0 / 6. This was below the eight-vendor average of 2.3 / 6.
- Channel-Level Optimization with MCP: 4.5 / 8. This was below the eight-vendor average of 5.2 / 8.
6. Fospha MCP (research score: 24.3 out of 48)

What is Fospha MCP?
Fospha MCP connects Fospha's measurement data to AI tools, so teams can ask plain-language questions using the same data as in the Fospha dashboard. Fospha announced its MCP server on April 9, 2026, and its setup guide covers Claude and ChatGPT.
Best for, based on the scores in this research: teams whose priorities are MCP interoperability and day-to-day marketing data reporting, which are Fospha's highest-scoring categories relative to the available points.
Fospha received an average research score of 24.3 out of 48, the sixth-highest total in this eight-vendor comparison.
Public documentation cited in this research: Fospha MCP setup and onboarding guide, Fospha FAQ on AI tools. Criterion-level evidence for Fospha is in the full research notes.
Fospha: research score by category
| Category | Fospha | Average research score across vendors | Highest research score among the vendors |
|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 4.3 / 6 | 4.7 / 6 | 6.0 / 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 2.4 / 5 | 3.3 / 5 | 5.0 / 5 |
| 3. Channel-Level Optimization with MCP | 4.6 / 8 | 5.2 / 8 | 7.9 / 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 3.3 / 8 | 3.0 / 8 | 5.0 / 8 |
| 5. Incrementality Testing with MCP | 1.3 / 5 | 2.4 / 5 | 3.6 / 5 |
| 6. MCP interoperability and workflows | 3.8 / 5 | 3.4 / 5 | 4.3 / 5 |
| 7. Analytical Backbone | 2.0 / 5 | 2.9 / 5 | 3.8 / 5 |
| 8. Enterprise-Grade Platform | 2.8 / 6 | 2.3 / 6 | 4.5 / 6 |
| Total score out of 48 | 24.3 / 48 | 27.2 / 48 | 38.3 / 48 |
Percentage of available points
Bars show the score as a percentage of the category maximum.
Fospha: highest-scoring categories
- MCP interoperability and workflows: 3.8 / 5. This was above the eight-vendor average of 3.4 / 5. Within this category, its highest-scoring criteria included documenting connections with Claude and ChatGPT, exposing tool and metric documentation, and returning reusable structured data.
- Marketing Data Reporting via MCP: 4.3 / 6. This was below the eight-vendor average of 4.7 / 6. Within this category, its highest-scoring criteria included reporting online sales progress, reporting digital media data, and reporting business outcomes beyond revenue.
- Channel-Level Optimization with MCP: 4.6 / 8. This was below the eight-vendor average of 5.2 / 8. Within this category, its highest-scoring criteria included simulating a user-specified budget change, reporting marginal returns and response curves, and forecasting a media plan.
Fospha: lowest-scoring categories
- Incrementality Testing with MCP: 1.3 / 5. This was below the eight-vendor average of 2.4 / 5.
- Analytical Backbone: 2.0 / 5. This was below the eight-vendor average of 2.9 / 5.
- Campaign & Ad Set-Level Optimization with MCP: 3.3 / 8. Despite being one of Fospha’s relatively lower-scoring categories, its score was still above the eight-vendor average of 3.0 / 8.
7. Haus (research score: 18.9 out of 48)

About Haus
Haus is an incrementality testing platform. Haus states that an MCP makes its experiment, MMM, causal attribution, KPI and spend data available to a brand's preferred LLM tools. This research found no public MCP documentation for Haus, which limited the scores it could receive.
Best for, based on the scores in this research: teams whose main requirement is incrementality testing, which is the category where Haus scored highest relative to the other vendors (tied third with Sellforte).
Haus received an average research score of 18.9 out of 48, the seventh-highest total in this eight-vendor comparison.
Public documentation cited in this research: What is agentic marketing attribution?, Haus Causal MMM. Criterion-level evidence for Haus is in the full research notes.
Haus: research score by category
| Category | Haus | Average research score across vendors | Highest research score among the vendors |
|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 3.1 / 6 | 4.7 / 6 | 6.0 / 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 1.9 / 5 | 3.3 / 5 | 5.0 / 5 |
| 3. Channel-Level Optimization with MCP | 2.9 / 8 | 5.2 / 8 | 7.9 / 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 3.1 / 8 | 3.0 / 8 | 5.0 / 8 |
| 5. Incrementality Testing with MCP | 2.6 / 5 | 2.4 / 5 | 3.6 / 5 |
| 6. MCP interoperability and workflows | 1.3 / 5 | 3.4 / 5 | 4.3 / 5 |
| 7. Analytical Backbone | 2.3 / 5 | 2.9 / 5 | 3.8 / 5 |
| 8. Enterprise-Grade Platform | 1.8 / 6 | 2.3 / 6 | 4.5 / 6 |
| Total score out of 48 | 18.9 / 48 | 27.2 / 48 | 38.3 / 48 |
Percentage of available points
Bars show the score as a percentage of the category maximum.
Haus: highest-scoring categories
- Incrementality Testing with MCP: 2.6 / 5. This was above the eight-vendor average of 2.4 / 5. Within this category, its highest-scoring criteria included reporting geo-test results, reporting own-media A/B test results, and prioritizing experiments.
- Marketing Data Reporting via MCP: 3.1 / 6. This was below the eight-vendor average of 4.7 / 6. Within this category, its highest-scoring criteria included reporting digital media data, reporting business outcomes beyond revenue, and reporting online sales progress.
- Analytical Backbone: 2.3 / 5. This was below the eight-vendor average of 2.9 / 5. Within this category, its highest-scoring criteria included calibrating the underlying MMM with incrementality tests, providing deterministic answers backed by Bayesian MMM, and reporting model validation and modelling KPIs.
Haus: lowest-scoring categories
- MCP interoperability and workflows: 1.3 / 5. This was below the eight-vendor average of 3.4 / 5.
- Enterprise-Grade Platform: 1.8 / 6. This was below the eight-vendor average of 2.3 / 6.
- Channel-Level Optimization with MCP: 2.9 / 8. This was below the eight-vendor average of 5.2 / 8.
8. Northbeam MCP (research score: 17.6 out of 48)

What is Northbeam MCP?
Northbeam MCP is a read-only connection that lets an AI assistant query a customer's Northbeam dashboards and summarize the results, following the user's existing Northbeam permissions. Northbeam's documentation covers Claude and ChatGPT, and states that the MCP is available on every Northbeam plan.
Best for, based on the scores in this research: teams that mainly need marketing data reporting in an AI assistant, which is Northbeam's highest-scoring category.
Northbeam received an average research score of 17.6 out of 48, the eighth-highest total in this eight-vendor comparison.
Public documentation cited in this research: Connect Northbeam to your AI assistant, Northbeam MCP server FAQs. Criterion-level evidence for Northbeam is in the full research notes.
Northbeam: research score by category
| Category | Northbeam | Average research score across vendors | Highest research score among the vendors |
|---|---|---|---|
| 1. Marketing Data Reporting via MCP | 5.0 / 6 | 4.7 / 6 | 6.0 / 6 |
| 2. Historical Performance Insights & Causal Explanation via MCP | 1.6 / 5 | 3.3 / 5 | 5.0 / 5 |
| 3. Channel-Level Optimization with MCP | 2.3 / 8 | 5.2 / 8 | 7.9 / 8 |
| 4. Campaign & Ad Set-Level Optimization with MCP | 1.4 / 8 | 3.0 / 8 | 5.0 / 8 |
| 5. Incrementality Testing with MCP | 1.1 / 5 | 2.4 / 5 | 3.6 / 5 |
| 6. MCP interoperability and workflows | 3.5 / 5 | 3.4 / 5 | 4.3 / 5 |
| 7. Analytical Backbone | 1.5 / 5 | 2.9 / 5 | 3.8 / 5 |
| 8. Enterprise-Grade Platform | 1.3 / 6 | 2.3 / 6 | 4.5 / 6 |
| Total score out of 48 | 17.6 / 48 | 27.2 / 48 | 38.3 / 48 |
Percentage of available points
Bars show the score as a percentage of the category maximum.
Northbeam: highest-scoring categories
- Marketing Data Reporting via MCP: 5.0 / 6. This was above the eight-vendor average of 4.7 / 6. Within this category, its highest-scoring criteria included reporting online sales progress, reporting digital media data, and reporting business outcomes beyond revenue.
- MCP interoperability and workflows: 3.5 / 5. This was above the eight-vendor average of 3.4 / 5. Within this category, its highest-scoring criteria included documenting connections with Claude and ChatGPT, returning reusable structured data, and exposing tool and metric documentation.
- Historical Performance Insights & Causal Explanation via MCP: 1.6 / 5. This was below the eight-vendor average of 3.3 / 5. Within this category, its highest-scoring criteria included reporting incremental ROAS and revenue for digital channels, reporting promotion-driven revenue, and reporting incremental ROAS and revenue for offline channels.
Northbeam: lowest-scoring categories
- Campaign & Ad Set-Level Optimization with MCP: 1.4 / 8. This was below the eight-vendor average of 3.0 / 8.
- Enterprise-Grade Platform: 1.3 / 6. This was below the eight-vendor average of 2.3 / 6.
- Incrementality Testing with MCP: 1.1 / 5. This was below the eight-vendor average of 2.4 / 5.
Frequently Asked Questions
1. What is the best MCP server for MMM?
Sellforte MCP received the highest total research score in this comparison of eight vendor MCP servers for MMM and incrementality testing. The ranking by total score is Sellforte (38.3 out of 48), followed by Recast (31.4), then Triple Whale (30.6). The scores are based on public documentation, so use them as a starting point and confirm the workflows you need in a demo.
2. What is the best MCP server for incrementality testing?
For incrementality testing through MCP, the highest research scores went to Recast (3.6 out of 5), followed by Measured (3.0), then Sellforte and Haus (2.6 each). On the three criteria for reporting test results (geo tests, own-media A/B tests and Meta Conversion Lift tests), Recast and Sellforte received the highest combined scores. Recast’s evidence includes retrieving imported experiment summaries. Measured received full credit for prioritizing experiments. The category separates result retrieval, prioritization and test design, so a single total does not establish end-to-end test execution.
3. Which vendors offer an MCP server for MMM and incrementality testing?
This research covers eight vendors: Sellforte, Recast, Triple Whale, Lifesight, Measured, Fospha, Haus and Northbeam. Seven of them publish MCP documentation. Haus describes an MCP in its marketing material, and no public MCP documentation was found for it. Open-source MCP servers such as mcp-server-meridian and simba-mcp are an option for teams that run their own models.
4. Does Sellforte have an MCP server?
Yes. Sellforte MCP connects all Sellforte product modules to AI assistants that support MCP, including Claude, ChatGPT and Gemini. It covers Marketing Mix Modeling results, incrementality tests, budget optimization and scenario planning, and campaign-level recommendations.
5. Are there open-source MCP servers for MMM?
Yes. Examples include mcp-server-meridian, published by BlueAlpha for fitted Google Meridian models, and simba-mcp for Simba's Bayesian MMM platform. They were not scored in this research. See open-source and self-built MCP servers for MMM.
6. Which vendor scored highest for channel planning?
Recast received 7.9 out of 8, followed by Sellforte at 7.4. Recast’s result includes documented optimizer and forecaster APIs.
7. Does connecting an AI assistant make an MMM result causal or more accurate?
The connection gives the assistant access to exposed data and tools. The credibility of the result still depends on the underlying model, experimental evidence, assumptions, validation and appropriate interpretation. The research does not test model accuracy or causal validity in production.
8. How should a team use this comparison to create a shortlist?
Start with the workflow that matters to your team, then inspect the relevant criterion scores and source documents. Ask each vendor to demonstrate that workflow using the intended client and access route. Verify returned metrics, supported granularity, model freshness, permissions and any steps that require the platform UI.
Change Log
| Date | Change |
|---|---|
| September 24, 2026 | This research article was published, including evaluation criteria and evaluation of eight vendors. |
| November 15, 2026 | Restructured the article: added a short answer summary, vendor descriptions with links to public documentation, an overview of open-source options and a question-led FAQ, and consolidated the scoring notes. Scores, vendors and evaluation dates are unchanged. |
Evaluation Dates by Vendor and LLM
Model names and evaluation dates below are recorded in the supplied evaluation log. There are 16 assessments: two per vendor.
| Vendor | LLM | LLM version | Evaluation date |
|---|---|---|---|
| Fospha | ChatGPT | GPT-6 Astra | September 23, 2026 |
| Fospha | Claude | Opus 5.5 | September 23, 2026 |
| Haus | ChatGPT | GPT-6 Astra | September 23, 2026 |
| Haus | Claude | Opus 5.5 | September 23, 2026 |
| Lifesight | ChatGPT | GPT-6 Astra | September 23, 2026 |
| Lifesight | Claude | Opus 5.5 | September 23, 2026 |
| Measured | ChatGPT | GPT-6 Astra | September 23, 2026 |
| Measured | Claude | Opus 5.5 | September 23, 2026 |
| Northbeam | ChatGPT | GPT-6 Astra | September 23, 2026 |
| Northbeam | Claude | Opus 5.5 | September 23, 2026 |
| Recast | ChatGPT | GPT-6 Astra | September 23, 2026 |
| Recast | Claude | Opus 5.5 | September 23, 2026 |
| Sellforte | ChatGPT | GPT-6 Astra | September 23, 2026 |
| Sellforte | Claude | Opus 5.5 | September 23, 2026 |
| Triple Whale | ChatGPT | GPT-6 Astra | September 23, 2026 |
| Triple Whale | Claude | Opus 5.5 | September 23, 2026 |
Limitations and Disclosures
Publisher affiliation. Sellforte designed and publishes this framework and is one of the vendors assessed. Claude and ChatGPT assigned the recorded scores. The article’s interpretation is editorial commentary based on those results.
Evidence scope. The evaluation uses vendor-owned public materials and technical documentation. It is not a hands-on test and does not measure pricing, implementation quality, support quality, performance, statistical accuracy or customer outcomes. Missing evidence is not proof of a missing capability.
Snapshot. The recorded evaluations are dated September 23, 2026. Product access and documentation can change. The scores reproduce the supplied workbook rather than silently incorporating new scores during article preparation.
Corrections. Vendors can submit documentation for correction or re-evaluation to research@sellforte.com.
Recommendation for buyers. Use this comparison as a structured starting point for your own evaluation, not as a final answer. We strongly recommend conducting evaluation calls or demos with vendors to verify fit against your organization's specific requirements.
Further Reading and Resources
Explore related guides, vendor comparisons and research on MCP, AI, marketing measurement and media optimization.
MCP, AI and Agentic MMM
- Introducing Sellforte MCP: Bring Sellforte Into Your AI Workflows
- The Rise of Agentic MMM (Marketing Mix Modeling): How AI Is Transforming Media Optimization
- Webinar: Agentic MMM in Action: The Future of Autonomous Media Planning and Buying in Real Time
- What is the Model Context Protocol (MCP)? — Official documentation
- Budget optimization & Scenario planning with Sellforte MCP (Channel-level)
- Marketing Reporting with Sellforte MCP
- Incrementality testing with Sellforte MCP
Vendor Comparisons and Evaluation Guides
- MCP Server Evaluation Checklist for MMM and Incrementality Testing: 48 Criteria
- 6 Best Conversational AI Tools for MMM and Incrementality Testing in 2026: An In-Depth Comparison
- Best Incrementality Testing Tools in 2026: In-depth Vendor Comparison
- Best MMM Solutions for Retail Brands: Top Marketing Mix Modeling Platforms (2026 Q3 update)
- Best MMM Tools for Ecommerce Brands: Top 10 Software for 2026
- How to Choose a Conversational AI Tool for MMM and Incrementality Testing: 48 Evaluation Criteria
MMM and Incrementality Measurement Fundamentals
- What is Marketing Mix Modeling (MMM)? A Complete Guide for Marketers
- What is incrementality testing? A practical guide
- What is Causal Marketing Mix Modeling (MMM)?
- What does "Enterprise-Grade" Mean in Marketing Mix Modeling (MMM)?
- Calibrating MTA with Incrementality: From Attributed ROAS to Incremental ROAS
Experiment Design and MMM Calibration
- Calibrating Marketing Mix Models with Experiments and Attribution data
- How to Integrate Experiments Into an MMM Platform: A Practical Guide
- How should we prioritize incrementality tests across markets and channels?
- How should we weight multiple incrementality tests when calibrating an MMM?
- MMM Calibration Explained: What It Is, Why It Matters, and How to Do It
Media Planning and Campaign Optimization
- ROAS, iROAS, miROAS: Choosing the Right KPI for Optimizing Media Spend
- Advertising response curves: What are they and why do you need them?
- Bid Optimization: How to Calculate the Optimal Bid Values for Your Campaigns Using miROAS
- The Shift in Marketing Mix Modeling: Why Campaign-Level Optimization is Taking Over
- The Five Requirements of Autonomous Media Buying and Optimization: M.A.G.I.C.
Original Research by Sellforte Labs
Author

Lauri Potka is the Chief Operating Officer at Sellforte, with over 15 years of experience in Marketing Mix Modeling, marketing measurement and media spend optimization. Before joining Sellforte, he worked as a management consultant at the Boston Consulting Group, advising some of the world’s largest advertisers on marketing optimization. Follow Lauri on LinkedIn.
You May Also Like
These Related Stories

MCP Server Evaluation Checklist for MMM and Incrementality Testing: 48 Criteria
6 Best Conversational AI Tools for MMM and Incrementality Testing in 2026: An In-Depth Comparison


