<aside> 🟢
+10% Revenue Per Visitor in 90 days. Or we refund the engagement.
</aside>
Every analytics tool you pay for ships a dashboard somebody else decided you needed. The schema is fixed, the metric definitions are theirs, and the day you ask a question that doesn't fit the tiles you export to a spreadsheet and rebuild it by hand at 11pm.
You don't have to. Point a reasoning model at your raw store export and there is no fixed schema. It answers the weird question in one prompt, in your definitions, ranked the way you actually make decisions. Below are four metrics no vendor will pre-build for you, each with the exact prompt you can run on your own export tomorrow.
You want contribution margin per customer for first-order, Meta-acquired customers, only on orders containing your hero SKU, in the 14 days before a price change versus the 14 days after. That is four filters stacked on a margin calc with a before-and-after split. No dashboard tile holds it, so today you export and burn an evening.
One prompt does it. Fill the brackets with your own definitions and run it against your Shopify order export.
<aside> ⚡
Steal this: Run this against your last price change on your top SKU. If per-customer contribution margin held or rose after the increase, you left pricing power on the table. Raise again.
</aside>
Blended MER fell 0.4 last week. Your dashboard tells you it dropped. It almost never tells you why in words you can repeat in a standup. MER is total revenue over total marketing spend, attribution-independent, and a drop has only a few possible drivers. The model can walk each one and attribute the change.
<aside> ⚡
Steal this: Next time a blended number moves and your team shrugs, run this before the meeting. Walk in with the single largest driver named and quantified instead of "we're looking into it."
</aside>
What is the true new-customer contribution of your Klaviyo welcome flow versus paid, and are you double-counting the same orders across both? Klaviyo knows flow-attributed revenue. Shopify knows the order and whether the customer is new. The ad platforms know spend. Most tools show you each in its own tab and leave the join to you.
The hard part is overlap. One email flow claims an order, a paid channel claims the same order, you add them, and you have invented revenue. This prompt forces the join and surfaces the contested orders instead of papering over them.
<aside> ⚡
Steal this: Run this on last quarter. The contested row is the number your CFO has never been shown. If it's large, your welcome-flow ROI and your paid CAC are both wrong in the same direction.
</aside>
Give me contribution margin by acquisition channel using my definition, ranked by margin dollars, and tell me which channels are actually losing money. Every brand defines contribution margin differently. Some stop at gross margin minus shipping. Some run it all the way through returns, processing, and acquisition spend. A tool picks one and you live with it. Here the definition is an input, so the answer is yours and it comes out the same every run.