Free tool · about 10 minutes

The pLTV Audit

Three read-only checks on your campaigns, revenue and data. They tell you whether a pLTV pilot will pay off.

The pLTV Audit is a free tool by Churney that checks whether predicted lifetime value (pLTV) bidding will work for your app or ecommerce business. It runs three checks on your own data: your ad campaigns, your revenue and your warehouse. Each check gives a plain verdict, the numbers behind it, and a full report for your team.

  • Free and self-serve, at tools.churney.io.
  • About ten minutes for all three checks.
  • Read-only. Nothing in your ad accounts or warehouse is changed.
  • Nothing is kept unless you press Save. Every check also runs on mock data, with no sign-in.

Updated November 2026.

Who it is for

Teams that buy users or customers on Meta, Google Ads or TikTok and wonder whether bidding on predicted long-term value would beat bidding on early events like installs, trials or first purchases. Subscription apps, mobile games, web-to-app funnels and ecommerce.

The three checks

Campaigns, opportunity, data.

CheckThe questionWhat it readsTimeVerdicts
1. Campaigns checkCan your campaigns carry a fair test?The last 28 days of your Meta, Google Ads or TikTok account, read-only1 minuteReady for a pilot · Ready after a few changes · Needs more volume first
2. Opportunity checkAre ad platforms bidding on the wrong customers?Revenue per customer over time, analysed in your browser2 minutesVery strong fit · Strong fit · Some headroom · Little headroom
3. Data checkCan your warehouse carry pLTV?Table and column names, plus read-only counts. Never rows5 minutesNo structural blockers · Not yet · No · Incomplete

1. Campaigns check: can your campaigns carry a fair test?

You sign in to Meta, Google Ads or TikTok with a read-only login. The check reads the last 28 days of campaigns, ad sets and conversions, and returns the campaign to test, the daily budget per test version and how many weeks the test takes.

How it is scored. The check looks for a campaign that can carry a split test at today’s spend, then rates the account on:

  • Volume. Optimisation events per week, against what a test cell needs to learn. Meta’s own bar for pLTV value optimisation is at least 100 conversion events a week for each of the last four weeks.
  • A test candidate. One established campaign that can be split into two versions at a budget the account can carry.
  • A built-in experiment. Whether the platform can run a native A/B or split test on that campaign type, or whether the test needs a geo split.
  • Structure. How much spend sits in campaigns optimised for conversions or value, how much is learning-limited, how stable the campaigns have been, bid caps, how broad the targeting is, and how many primary conversion actions compete.

The verdict:

  • Ready for a pilot. One campaign can carry a fair test and nothing in the structure gets in the way.
  • Ready after a few changes. A fair test fits your spend, but one check fails or two or more need attention. The changes are listed.
  • Needs more volume first. No campaign gets enough optimisation events a week for a test to show a difference.

A test plan is about six weeks: two weeks for the platform to learn, then four weeks of measurement.

2. Opportunity check: are ad platforms bidding on the wrong customers?

Ad platforms judge a new customer by what happens in the first days. If your best customers look ordinary on day 7, the platforms underbid for them, and a value signal has room to help. This check measures that gap.

You bring revenue per customer over time: a CSV, Stripe, Paddle, Shopify, Chargebee, AppsFlyer or a BigQuery table. The analysis runs in your browser. Customer rows never leave it.

How it is scored. The score is the Spearman rank correlation between each paying customer’s revenue at day 7 and at day 90 (day 60 or day 30 if your history is shorter). It needs at least 30 paying customers with that much history. A lower score means early revenue ranks your customers poorly, so there is more for a pLTV signal to fix.

ScoreVerdict
Below 0.70Very strong fit
0.70 to 0.85Strong fit
0.85 to 0.90Some headroom
0.90 and aboveLittle headroom

The result also shows what share of your best 10% of customers by day 90 were not in your top 10% on day 7. Those are the customers the platforms pass over.

3. Data check: can your warehouse carry pLTV?

A pLTV model needs user-level revenue, a stable user ID, enough history and data that does not change after the fact. This check reads table and column names and runs a handful of read-only count queries. It never reads values in your rows.

It connects to BigQuery directly. For Snowflake, Redshift, Databricks, Postgres or Athena, you copy the queries, run them yourself and upload the results. You also answer a few questions metadata cannot show, such as how refunds are recorded.

How it is scored. Four core areas decide the verdict:

  • Grain. Revenue and events are per user, not aggregated by day or campaign.
  • Outcomes. Revenue is tied to people, with enough months of history to learn from.
  • Identity. Revenue joins to the events of the same users.
  • Point in time. Rows are appended, not overwritten, so the model sees what was known at the time. Refunds arrive as their own rows.

Features, delivery to the ad platforms, attribution and operations are checked too. They become constraints to plan around, not blockers.

The verdict:

  • No structural blockers visible. Nothing in the core areas stands in the way. A data scientist still reviews it before anything is promised. Constraints, if any, are listed.
  • Not yet. A core area fails. The one fix that matters first is named.
  • No. There is no user-level revenue, or events are only stored in aggregate. That is structural.
  • Incomplete. A core question is still open, or a check did not run. The review says which.
  • Check the opportunity first. If most of a customer’s first-year revenue arrives in week one, early revenue may already be a good enough signal, and the review points to the Opportunity check.

After all three checks

With all three checks in, the audit gives:

  • A pilot plan sized on your numbers: the platform, the campaign and the budget.
  • Cohort-pLTV model recommendations for your data team: the cohort model to try first, with the settings filled in from what the checks measured. It is a starting point for your own cohort modelling.

What the audit does not do

Plainly:

  • It does not change anything in your ad accounts or warehouse.
  • It does not read values in your warehouse rows, and customer rows from the Opportunity check stay in your browser.
  • It does not prove an uplift. Only a split test against your current campaigns does that.
  • It does not check whether creative, pricing or product changes would do more for you than a value signal.

FAQ

Questions we get.

Is there a free tool to check if pLTV will work for my app?

Yes. The pLTV Audit at tools.churney.io is free and self-serve. It checks your ad campaigns, how well early revenue predicts long-term revenue, and whether your warehouse can carry a pLTV model. It takes about ten minutes.

How do I know if value-based bidding is worth it?

Measure how well early revenue ranks your customers. If your best customers by day 90 often looked ordinary on day 7, platforms bidding on early events pass them over, and predicted value has room to help. The Opportunity check measures this on your own revenue.

What data do I need for pLTV bidding?

User-level revenue history in a warehouse, a stable user ID on events and payments, enough months of history, and rows that are appended rather than overwritten. The Data check tests this from metadata.

Do I need to sign in?

Every check runs on mock data without signing in. To run on your own data you sign in with a Google work account.

Is my data stored?

No, unless you press Save. Saved results are kept 90 days.

Who makes it?

Churney, a pLTV bidding company. Churney predicts each new customer’s long-term value from first-party data and sends it to Meta, Google Ads and TikTok as the bidding signal.

Run the pLTV Audit.

Free, about ten minutes, read-only.

Start at tools.churney.io