Which channels have a steady ROI?

4 min read
Published Nov 18, 2026

Welcome to the Sellforte MCP Use Case Library. Each post takes one question that marketing teams really ask their AI assistant through Sellforte MCP, which is now in general availability and accessible to all marketers worldwide. It strips out anything customer-specific and walks through what the assistant does with it, step by step. We run every question on our demo account so the numbers you see are illustrative, not anyone's real results. The idea is simple: steal the prompts, try them on your own model, and see what you get back. We hope you find this series inspiring!

-Juha, CEO at Sellforte

Can you trust a channel's average ROI?

An average ROI is a handy number for planning, until you notice that one channel hits it every month and another only gets there because of one great month a year ago. Analysts who feed ROI into budgets want to know which averages are solid and which are hiding a lot of movement. That's what this question is for.

What you ask

Using the MMM's incremental ROI, not ad-platform ROAS, give me by ad platform from 2025-11 to 2026-10 the number of months with spend, the median monthly ROI, the min–max range, the standard deviation and total spend. Leave out months without spend.

Email and direct mail have the widest spread. Show their monthly ROI next to spend: is it noise or a trend?

What Sellforte MCP does

Naming the metric matters here. If you just say "ROAS", an assistant can easily reach for the ad platform's own figure, and then channels like email or TV show up as zero because no platform reports revenue for them. Asking for the MMM's incremental ROI avoids that.

The assistant pulls spend, incremental sales and ROI by ad platform and month for the 12 months, drops the months without spend, and calculates the statistics itself. Here's what came back on our demo account, sorted by spend:

Ad platformMonths with spendMedian monthly ROIMin–maxStd. dev.Spend
Google Ads124.383.41–4.850.47$7 394 532
Meta Ads125.973.41–7.611.16$2 825 945
Direct Mail113.561.34–6.031.51$2 021 429
Email121.451.02–5.841.39$1 735 959
TV61.361.08–2.060.34$1 690 985
Adtraction122.161.92–2.760.23$689 705
TikTok Ads124.573.25–5.230.74$390 017
Criteo122.501.81–3.600.54$369 601
Print93.072.82–3.380.17$159 373
Microsoft Advertising122.491.47–4.200.76$127 084
Awin123.512.16–5.040.97$88 946
PriceRunner124.033.20–4.830.57$12 934

 

Illustrative demo data, 2025-11 to 2026-10.

Pinterest Ads had spend in only one month, so the assistant leaves it out of the statistics instead of pretending one data point has a spread.

A few patterns jump out. Google Ads, the biggest channel by far, is also one of the steadiest: every month sits between 3.41 and 4.85. Adtraction and print are even tighter, just at lower or smaller levels. Meta has the highest median, but it moves around more, from 3.41 to 7.61. And the two channels with the widest spread relative to their median are email and direct mail.

The follow-up shows those two spreads mean very different things:

  • Email isn't noisy, it shifted. It ran at 5.84 in November 2025, 3.29 in December and 2.00 in January, then stayed between 1.02 and 1.67 every month from February to October 2026, on similar spend of roughly $120 000 to $155 000 a month. The median of 1.45 is the better guide for planning; the 12-month average is pulled up by last winter.
  • Direct mail swings with the size of the drop. Its two weakest months were also its two smallest: $14 286 of spend in August 2026 at an ROI of 1.34, and $57 143 in October 2026 at 2.07. Larger drops mostly landed between 2.96 and 6.03.

One caveat the assistant states clearly: the median, range and standard deviation are its own calculations over monthly results. They describe how results vary month to month, not the model's statistical uncertainty.

Why it matters

If you plan with averages, you're betting that next month will look like the average month. For Google Ads on the demo data, that's a reasonable bet. For email, the average describes a channel that no longer exists, and for direct mail the answer depends on how big the next drop is.

A volatile result deserves a second look before anyone moves budget on it. This table is a cheap way to find out where that second look is needed, and it takes one prompt.

Try these follow-ups

  • Do the same table by advertising channel within Meta Ads, and flag channels with fewer than 6 months of spend.
  • For Meta Ads, show monthly ROI next to monthly spend. Does ROI drop in the months with more spend?
  • Show the total monthly ROI trend across all channels for the same 12 months.
  • Which channels have incrementality experiments we could use to check the median ROI?

Related

Try it yourself

Want this table for your own channels? Book a demo, or check how to connect Sellforte MCP to your LLM client. The Sellforte MCP launch post explains the bigger picture.

We keep adding new use cases here, so check back from time to time.