Which ads bring paying customers?
Datarithms ranks your Google, Meta, TikTok and LinkedIn campaigns by what a paying customer actually costs, and tells you each week which stay up, which go down, and which are too early to judge.
| Pos. | Campaign | Spend, 28 days | Paying customers | Cost per paying customer | Verdict |
|---|---|---|---|---|---|
| 1 | Search, brand termsGoogle Ads | $1,240 | 31 | $40 | Keep |
| 2 | Reels, free trial offerMeta Ads | $2,910 | 34 | $86 | Keep |
| 3 | Lookalike, sign-upsMeta AdsCosts 2.9 times your best campaign per paying customer. | $5,570 | 48 | $116 | Cut 20% |
| – | Job title targetingLinkedIn Ads | $1,960 | 4 | Not shown | Too few to tell |
| – | Spark ads, demo videoTikTok Ads | $640 | None yet | Not shown | Too few to tell |
- Stays up: keep
- Goes down: cut or move budget
- Fewer than 30 paying customers: no position yet
Cost per paying customer = spend ÷ paying customers. Not shown below 30 paying customers.
From the ad to the paid line
Example dataPlatforms usually sort campaigns by spend or by their own conversions. The table sorts them by paying customers. Same five campaigns, different order.
Why your ad dashboards disagree with your bank account
Each ad platform reports the conversions it can see, counted by its own rules. None of them knows who actually paid you.
Every platform counts itself in
Google, Meta, TikTok and LinkedIn can all claim the same customer. Added together, they report more customers than you have. Why platform conversions add up.
A cheap sign-up is not a cheap customer
Campaigns tuned for sign-ups can fill your trial list with people who never pay. The platform reports that as success.
Nobody checks the change
You cut a budget and move on. Four weeks later nobody knows whether paying customers went up or down.
How the table is made
Three steps, then one short table each week.
You share ad and payment data
Today: export files from your ad platforms, plus Stripe or HubSpot for who paid. Direct read-only connections are being built.
We match each paying customer to an ad click
Last ad click within 90 days. Refunds within 14 days do not count. Customers we cannot match are listed as unmatched, never guessed.
You decide every change
Keep, cut, or too few to tell, with the numbers behind each. Nothing changes in your ad accounts without your approval, and each change is re-checked after 28 days.
Four ad platforms, one table
What each platform reports, and what we need from it today.
| Platform | What it reports | What we need today |
|---|---|---|
| Google Ads | Clicks, cost and the conversions its tag records | Campaign cost by day, as an export file |
| Meta Ads | Results inside its attribution setting, including some views | Campaign cost by day, as an export file |
| TikTok Ads | Conversions inside its attribution window, including some views | Campaign cost by day, as an export file |
| LinkedIn Ads | Leads and conversions it can see on LinkedIn | Campaign cost by day, as an export file |
Who it is for
Software and service companies
You advertise for sign-ups, trials or leads, and the money arrives days or weeks later through Stripe or a sales pipeline.
Agencies
You run ads for clients and need a dated record of each change and what happened next, in paying customers, for the monthly review.
Questions before you ask for access
Is Datarithms a finished product?
No. It is in early access. For the first companies we prepare the weekly table together with you from your own export files, while direct read-only connections to the ad platforms are built.
Do you change my ads?
No. Datarithms proposes changes with the numbers behind them. Nothing changes in your ad accounts without your approval.
What does it cost?
Pricing is not set yet. It is one of the open decisions during early access, so ask us when you request access.
Why is there no verdict for some campaigns?
A campaign with fewer than 30 paying customers gets no position and no verdict. We show its counts and say it is too few to tell.
Request early access
Tell us what you run. We reply by email within two working days.
Early access: we prepare your weekly table with you from your own exports while the direct connections are built.