Signal Studio
Every purchase looks the same to Meta.
Not to you.
Your ad account counts a ₹499 cash-on-delivery order that will be refused at the door exactly the same as a ₹4,000 prepaid order from a customer who buys four times a year. It calls retargeting your best campaign because it sells to people you already own. It cannot see any of this, because nobody told it.
Questions this answers
- Am I paying prospecting prices to re-buy my own customers?
- Why is my ROAS healthy and my bank balance flat?
- How much of my spend trains Meta on orders that come back?
- What can I actually put in Ads Manager beside ROAS?
The diagnosis
Four ways a brand leaks money. You have one of them.
A list of eighteen signals is a decision, and decisions get deferred. So the product runs four tests against your own orders and names the one to fix first, with the campaigns behind it. Every share is counted only against the orders it could be counted against - if a third of your orders never carried a discount field, it says so rather than inventing a number.
- One answer, not a menu of eighteen switches
- Each test names the campaigns driving it
- What could not be judged is listed, not hidden
- Counted from the same nightly tables as your reports
Lead marked Qualified
Moved by a rep, or by a scoring rule
Match keys attached
QualifiedLead posted to Meta and Google Ads, deduplicated on your own lead id.
The harvest leak
You are paying prospecting prices to re-buy your own customers.
A purchase from a four-time customer counts as a win, so prospecting takes credit for harvesting demand that already existed. Send first orders as their own signal and the split becomes a column: a campaign at 0.3 new visitors per click is sending you people you already have, and that is the number retargeting cannot hide from.
- First purchases divided by all purchases, per campaign
- Cost per first customer - the real acquisition cost
- New visitors per click, beside Clicks
- Meta bids on exactly this split already
The RTO leak
A purchase you have already sent cannot be taken back.
Fire Purchase at checkout on a cash-on-delivery order and you have cast an irreversible vote for a customer who may never pay. Hold that event until the carrier confirms delivery, report the returns, and send prepaid orders as their own signal - because a browser closed days before the parcel arrived, and only your systems know what happened next.
- Purchase held until delivery is confirmed
- Returns and cancellations reported back
- Prepaid share and cost per delivered order
- The ROAS column starts agreeing with the bank
Third-party pixel
blocked
cdn.some-tracker.com/px.js
First-party path · your own domain
arrives
yourstore.com/a7f2k/collect
Attribution coverage
0%
of orders tied to a real source on live stores
The discount leak
The sale ends. The account keeps hunting discount buyers.
A discount-led push flatters its own ROAS with orders that were coming anyway, and quietly teaches the algorithm what kind of buyer to find next. Send full-price orders as a separate signal - before the sale, not during it - and you keep a clean line running while the noisy one spikes.
- Full-price share per campaign
- Discount, coupon code and cash discount on every order
- Switch it on before the sale, not during
- Cost per full-price order, beside ROAS
Scored here, not in Events Manager
Meta tells you the number weeks later, in aggregate. We score each event from the keys we attach, and name what to add.
The margin leak
Budget follows conversion rate. It should follow contribution.
Ad platforms optimise toward whatever converts cheapest, which is almost always your cheapest product rather than the one that pays for the business. Send margin and category on the order and spend can chase contribution instead - and where you hold no cost data, every report says so plainly rather than inventing a margin.
- One event per category, expanding as you add them
- Gross profit and margin on the order itself
- Cost per purchase by category, side by side
- No cost data? It says so, rather than guessing
In Ads Manager
Columns you cannot get anywhere else.
Each signal becomes a custom conversion, and each custom conversion becomes a column beside Clicks and ROAS. The product shows you the exact rule to build, whether the signal is switched on, and how many events have actually left - so you know the column will have something in it before you go looking.
- The exact custom-conversion rule, ready to build
- Whether it is on, and what has actually been sent
- 47 parameters to filter on, with real delivery rates
- Your existing ROAS column stays exactly where it is
Your pixel sends the order. This is what we know about it.
Built on one parameter, in Meta
Purchase where is_first_purchase = true → CB First-time buyers · cost per genuinely new customer, per campaign
Inside the module
Inside Signal Studio
One catalogue of everything we can send, and one honest answer about what you are sending today.
Where you lose money
Four tests, one thing to fix, the campaigns behind it
The catalogue
Every event, parameter and audience in one place
Would fire vs sent
The ceiling and the reality, side by side
Set-up route
Five steps from the finding to a working column
Coverage, stated
What could not be judged, and why
Delivery per parameter
What actually left, not what could have
Nothing here is a projection. Every figure is a counted row from the same nightly tables the rest of your reports read, so the diagnosis can never disagree with the page it sends you to.
The other modules
Attribution
Every order credited to the ad that earned it, with the receipt and a match score.
OpenGrowth intelligence
Command Center, Today’s Actions and What We Found: what to do on Monday, priced on your costs.
OpenAnalytics & Insights
Twenty reports on one date window, every one ending in a decision.
OpenAudiences
Segments from behaviour the store records, sized against what Meta can match before they sync.
OpenClose the loop on your next lead
Find out which leak you have
Measurement first, then everything else. Spin up your workspace in minutes and watch the first signal go back on your very first order. 30-day free trial, no card required.