Ask Sellforte MCP: How did last week compare with last year?
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
The Monday question: how did last week really go?
A week closes, and by Monday morning someone wants to know what happened. If you're a marketing manager or head of digital, you need a short readout you can paste into the weekly update: what moved in spend, incremental sales and ROI, and which channels drove it. It's the most common analysis people ask for, and it's usually more interesting against last year than against the week before.
What you ask
What can you tell me about last week? Compare it with the same week last year and list the best and worst channels.
Now put the same week next to the week before. Did anything move?
What Sellforte MCP does
We ran both prompts on our demo account. The demo data runs to 2026-11-01 and weeks start on Sunday, so "last week" is 2026-10-25 to 2026-10-31, and the same week last year is 2025-10-26 to 2025-11-01. All figures below are illustrative demo data in USD.
Starts with the whole picture
Before looking at channels, the assistant split total sales into base, promotions, media and weather. Total sales came in at 3 013 593, down 15.05% from 3 547 458 a year earlier. That headline hides three different stories:
| Component | Same week 2025 | Last week | Change |
|---|---|---|---|
| Base | 1 374 588 | 1 446 326 | +5.22% |
| Promotions | 534 353 | 693 719 | +29.82% |
| Media | 1 628 680 | 873 509 | -46.37% |
| Weather | 9 837 | 38 | -99.61% |
So underlying demand and promotions were up. The drop came from media.
Then goes channel by channel
Next it pulled spend, incremental sales and ROI by media group for both weeks. Media spend fell 39.43%, from 367 688 to 222 717, and incremental sales from media fell 48.32%, from 1 583 377 to 818 238. Total media ROI went from 4.31 to 3.67.
The biggest single gap is Direct Mail. Last year it spent 108 571 in this week and returned 761 913 in incremental sales, an ROI of 7.02. This year it didn't run at all. Paid Video and Display were also off this year, while both had spend in the same week last year.
| Media group | Spend | Incremental sales | ROI | ROI same week 2025 |
|---|---|---|---|---|
| Cross-network | 88 115 | 399 660 | 4.54 | 4.64 |
| Paid Social | 56 392 | 257 400 | 4.56 | 2.59 |
| Paid Search | 18 636 | 66 047 | 3.54 | 3.13 |
| 35 833 | 37 335 | 1.04 | 0.59 | |
| Affiliates | 14 047 | 33 262 | 2.37 | 2.21 |
| Direct Mail | 0 | 0 | – | 7.02 |
From there the best and worst lists follow. Cross-network is the biggest volume driver at 399 660. Paid Social is the best improver: on about 9% more spend, its incremental sales nearly doubled (+92.37%) and ROI went from 2.59 to 4.56. At the bottom, Email has the lowest ROI of the channels that ran, at 1.04, and Paid Search lost 28.87% of its incremental sales after spend was cut by 37.25%.
And checks the week before
The follow-up tells a calmer story. Against 2026-10-18 to 2026-10-24, spend was up 2.58%, incremental sales up 3.05%, and ROI flat at 3.67 versus 3.66. Cross-network added 58 242 in incremental sales, while Paid Social gave back 23 091 and Paid Search 8 053. Week on week, nothing dramatic happened. The year-on-year gap was built earlier, when channels like Direct Mail stopped.
One detail worth knowing: the media figure in the sales split (873 509) and the channel total (818 238) differ on purpose. The split counts media's effect in the week it lands, so it adds up to total sales. The channel view credits carryover back to the week the money was spent, which is the basis for ROI. The assistant pointed this out rather than leaving you to wonder.
Why it matters
A weekly number on its own doesn't tell you what to do. Sales down 15% could mean weaker demand, a missing promotion or a channel you switched off, and each one needs a different response. Here the model shows that demand and promotions were actually up, and the gap is mostly media you chose not to run. That's a much better conversation to have with your team than "sales are down".
It also shows why the comparison matters. Week on week, everything looks stable. Against last year, there's a clear question: was stopping Direct Mail a decision you'd make again?
Try these follow-ups
- Rank the channels by ROI, but only those with at least 10 000 in spend last week.
- Show the last eight weeks of media spend and incremental sales as a weekly trend.
- Split last week's change by country. Is the drop the same everywhere?
- Which weeks last year did Direct Mail run, and what did it return?
- Do the same review for month-to-date against the same days last year.
Related
- Marketing mix modeling
- Causal attribution
- Ask Sellforte MCP: Why did our Direct Mail ROI drop last quarter?
- Sellforte MCP launch post
Try it yourself
Want to run this review on your own data? Book a demo, or if you're already a Sellforte customer, see how to connect Sellforte MCP to your LLM client. New use cases land in this library regularly.
