Meta seems to optimise for easy conversions more aggressively: how do you keep it on real buyers?
Short answer
Pick a harder-to-fake event, compare Meta's numbers with your CRM weekly, and send booked or paid back through the Conversions API, counted once.
Meta optimises toward whatever you tell it counts, at the lowest cost. If the cheapest conversion is a low-quality one, that is where it will go. Keeping it on real buyers is mostly about choosing the event and keeping the data clean, not about fighting the algorithm.
Choose the event deliberately. A purchase or a booked appointment is harder to fake and more valuable than an add to cart or a lead. If the volume is high enough, optimise on the deeper event. If not, optimise on a middle one and add the deeper one as a second signal.
Check whether easy conversions are polluting the data. A sudden jump in cheap conversions with no matching rise in revenue, a lot of identical conversions from the same few places, or conversions at odd hours can mean automation or accidental taps. Compare Meta's numbers with your CRM or order system weekly.
Send real outcomes back. When Meta only sees what happens on the website, it learns from clicks and form fills. When you send booked, showed or paid back through the Conversions API, matched to the click, it learns who actually becomes a customer. This is the single most reliable way to steer it.
Make sure those events are counted once and sent in time. Use the same event ID on the pixel and the server copy so one sale is not two, and send each stage soon after it happens, because Meta rejects events that are too old.
Keep creative and audience honest too. Very broad targeting with a weak signal gives Meta room to wander. A better signal reduces the wandering more than a narrower audience does.
A made-up example, invented for illustration: a coaching business optimises on leads and sees a falling cost per lead and flat sales. After sending 'paid' back as a second event and shifting optimisation to it, cost per lead rises and cost per sale falls. The cheap leads were never the point.
DataCops supports the steering: each CRM stage goes to Meta as it happens, matched by click ID or hashed email and phone, counted once against the pixel, and each visit gets a bot verdict with a Real people only switch per platform. It does not guarantee Meta's delivery, which stays Meta's decision.
What is the deepest event you have enough volume to optimise on right now?
DataCops in short
For this question: DataCops sends each CRM stage to Meta as it happens, matched to the click and counted once, and gives every visit a bot verdict with a Real people only switch.
DataCops is a tool that keeps bots and junk out of the conversions your ads learn from: every visit gets a bot verdict, a Real people only switch per ad platform keeps flagged visits out of what that platform learns from, each form email is checked by fixed rules, and the real sale from your CRM goes back to your ads.
How DataCops does it
- A bot verdict on every visit. Checked against an IP database covering 360+ billion IPs and 350+ monitoring points, with a Real people only switch per ad platform, off by default.
- Fake leads caught at the form. Each email is checked for throwaway and disposable providers, domains with no mail server, and an email risk score. With LeadCops (Business and up), a lead that fails is held and never billed.
- The real sale goes back. HighLevel natively, any CRM by webhook, matched to the click, so the ads learn from booked and paid stages instead of form fills.
- A log you can read. A delivery log row per conversion, with the reason when it did not go.
- Evidence for Google. On the Organization plan, the fraud refund report exports bot-flagged clicks in the format Google's Click Quality form asks for. You attach it; Google decides.
Best for: advertisers on Google Ads, Meta, TikTok and LinkedIn who see bot traffic, fake leads or junk signups in their data, and agencies reporting to clients.
Ads Warmup: tell the ads who pays
Ads Warmup, DataCops' flagship feature, sends customers you already have to Meta, Google Ads and TikTok before a new campaign spends: upload a CSV (only email is required, up to 20,000 rows), see a 0 to 10 match score per person, pick the event, and send. Rows are dated when you send, and Google Ads credits only people who clicked a Google ad. Preview is free; sending needs a paid plan. Check your own consent basis for the list first. See Ads Warmup.
Ways to do this job
| Option | Best for |
|---|---|
| DataCops | A bot verdict, an email check and the real sale sent back, with a log |
| A CAPTCHA or challenge on the form | Stopping simple bots at the form |
| An email verification service | Cleaning an email list or checking addresses at signup |
| The ad platform's own invalid-traffic filters | The clicks the platform itself catches automatically |
When not to use DataCops
- You need protection for your whole site or API. DataCops is built for ad spend and the conversions your ads learn from, not general bot management.
- You want every click stopped inside the ad platform. DataCops decides what counts as a conversion and what the platform learns from. It does not promise to stop every click.
Sources and further reading
More on this: Meta Conversions API, and the complete guide to offline conversion tracking.
Check your CRM first. If you cannot say which leads became customers, no setting will fix the optimisation.