Ad attribution by customer lifetime value

Stop paying for customers who never come back.

Day-1 ROAS tells you what you sold. Adfinity tells you what each ad is worth over a customer's life. Built for e-commerce and subscription brands running Meta, Google and TikTok ads.

Book a walkthrough Read the method

Early access. 14-day trial, with a 24-month historical backfill of your orders.

Cohort revenue per customer

Cumulative, by months since first order

Illustrative data
  • A Lookalike 1%
  • B Broad, video
  • C Free-shipping promo
Same three campaigns, sorted by day-1 ROAS
Campaign CPA Day-1 ROAS 180-day LTV LTV:CAC1
C Free-shipping promo $19 1.8x $51 ±6 2.7
B Broad, video $28 0.9x $85 ±10 3.0
A Lookalike 1% $30 0.7x $120 ±14 4.0

1 Predicted 180-day revenue per customer divided by CPA. Ranges are confidence ranges. All figures on this card are made up to show the layout.

Why day-1 numbers mislead

Two campaigns, the same $24 CPA.

Your ad dashboard sees them as equals. Six months later they are not.

Campaign X sells a discounted starter kit. People buy once and leave. Campaign Y sells a full-price subscription. The first order is smaller, but customers keep reordering.

Judged on day-1 ROAS, you scale X and pause Y. Judged on what each customer is worth after 180 days, you do the opposite. Adfinity puts that second number next to the first, on every campaign, ad set and creative.

Example only. Numbers are illustrative.

X Starter-kit promo

180-day LTV

$41

CPA
$24
Day-1 ROAS
1.3x
LTV:CAC
1.7

Y Subscription offer

180-day LTV

$163

CPA
$24
Day-1 ROAS
0.8x
LTV:CAC
6.8

Method

Four steps from order to bid.

No pixels to babysit and no black box. Each step says what goes in and what comes out.

  1. Connect

    Authorise your store (Shopify, WooCommerce, or a read-only connection to your own database) and your ad accounts (Meta, Google Ads, TikTok Ads). We import 24 months of order history and ad spend.

    Out A single timeline of orders, customers and spend.

  2. Match

    Orders are matched to campaigns on the server, using click parameters and order data you already own. You get a first-touch view and a data-driven view side by side. No device fingerprinting. Orders we cannot match are shown as unattributed, not guessed.

    Out Each customer linked to the campaign, ad set and creative that acquired them.

  3. Predict

    A cohort model fits repeat-purchase curves and projects 90, 180 and 365-day LTV for each campaign, ad set and creative. Every forecast carries a confidence band. The model needs at least 3 months of order history. With less, it says so and shows observed revenue only.

    Out Predicted LTV with a range, and CAC payback in days.

  4. Act

    Predicted value is sent back to Meta, Google Ads and TikTok Ads as conversion value, so bidding favours customers who stay. Each week you get budget recommendations with the evidence behind them. You approve every change.

    Out Value-based conversions, plus a weekly shortlist of budget moves.

Product

What is in the dashboard.

Seven working views. Each one answers a question a media buyer asks every week.

LTV by campaign, ad set and creative
Predicted 90, 180 and 365-day LTV with confidence ranges, next to CPA and day-1 ROAS. Sort by whichever you trust.
Cohort explorer
Revenue per customer by acquisition month, channel, first product or discount code. Spot a weak cohort before it shows up in the monthly report.
CAC payback
Days until a cohort's contribution margin covers what it cost to acquire, for each campaign.
Creative leaderboard
Rank ads by predicted LTV instead of click-through rate. Often the creative that brings the best customers is not the one with the most clicks.
Value-based conversion sync
Send predicted customer value to Meta, Google Ads and TikTok Ads as the conversion value. Review what was sent in a log.
Alerts
Plain-language rules, such as "CAC payback above 120 days on Campaign X". Delivered by email.
Exports
Google Sheets, BigQuery or CSV. Your numbers stay yours.

Metrics

Five terms, defined once.

The same definitions are used across the product and in the calculator below.

LTVa
Lifetime value: total revenue a customer brings over time. We report it for a fixed window (90, 180 or 365 days) so campaigns compare fairly.
CACb
Customer acquisition cost: ad spend divided by new customers in the same period. CPA counts any purchase; CAC counts only first-time buyers.
LTV:CACc
Contribution LTV divided by CAC. A ratio of 3 or more is a common rule of thumb. The right target depends on your margin, cash position and growth goals.
Payback periodd
Months until gross margin from a customer repays their CAC. It tells you how long your cash is tied up.
Contribution-margin LTVe
LTV after product cost, shipping and payment fees. Revenue LTV flatters low-margin products, so Adfinity uses margin wherever you give us a margin.

LTV:CAC calculator

A back-of-envelope check with your own numbers. Nothing is sent anywhere. It runs in your browser.

Contribution LTV
$165.00
LTV:CAC
3.7 : 1
Payback
6.5 months

At these inputs, each customer returns about 3.7 times what they cost to acquire.

LTV = AOV x margin x orders/yr x years
Payback (months) = CAC / (AOV x margin x orders/yr / 12)

  1. a Adfinity forecasts LTV per cohort and shows a range, not a single figure.
  2. b Blended CAC hides channel differences. Use CAC per campaign.
  3. c A ratio built on revenue instead of margin overstates the result.
  4. d Payback ignores what happens after it. Read it with LTV, not instead of it.
  5. e Margin inputs are yours. We do not estimate costs for you.

Data and privacy

Your customers' data stays yours.

An attribution tool sees a lot. These are the rules we build to, in plain terms. The legal detail is in the Privacy Policy.

  • First-party data only

    We use orders from your store and spend from your ad accounts. We do not buy or add third-party audience data.

  • Aggregated reporting

    Dashboards show cohorts and campaigns, not individual shoppers. Customer-level rows are used for matching and modelling only.

  • No cross-site tracking

    We do not follow people across other companies' sites or apps, and we do not build profiles outside your own store.

  • No fingerprinting

    Matching uses click parameters and order data. No canvas, font or device-signal fingerprints.

  • Deletion on request

    Email us and we delete your workspace data and the customer records we hold for you, then confirm in writing.

  • Regional hosting

    You choose where your workspace data is hosted when we set it up, and it stays in that region.

Pricing

Priced by the revenue you track.

No per-seat fees and no cut of ad spend. Every plan starts with a 14-day trial and a 24-month historical backfill.

Launch

$149/ month

Up to $250k / mo tracked revenue

  • LTV by campaign, ad set and creative
  • Cohort explorer and CAC payback
  • Value-based conversion sync
  • Alerts and CSV export
Book a walkthrough

Scale

$449/ month

Up to $2M / mo tracked revenue

  • Everything in Launch
  • Multiple stores and brands in one workspace
  • Creative leaderboard by LTV
  • Google Sheets and BigQuery export
Book a walkthrough

Custom

Let's talk

Above $2M / mo tracked revenue

  • Everything in Scale
  • Pricing set to your volume
  • Contract terms by agreement
  • Onboarding help with your data team
Book a walkthrough

Prices are in US dollars. Tracked revenue is the total order value we import from your store in a month.

FAQ

Straight answers.

Something missing? Write to support@adfinity.life.

Is Adfinity available today?

Adfinity is in early access. We onboard new brands through a walkthrough so we can check that your order history fits the model before you start the 14-day trial.

How accurate are the LTV predictions?

They are estimates, and every one is shown with a confidence range rather than a single number. The model needs at least 3 months of order history to forecast. More history narrows the range. Promotions, new products and unstable repeat-purchase behaviour widen it. As cohorts mature you can compare past forecasts with what customers actually spent.

How does Adfinity cope with iOS privacy changes and cookie loss?

It does not rely on device identifiers or fingerprinting. Orders are matched on the server side, using click parameters and order data you already own. When an order cannot be matched to a campaign, it is reported as unattributed instead of being guessed.

Which stores and ad platforms are supported?

Stores: Shopify, WooCommerce, or your own database. Ad platforms: Meta, Google Ads and TikTok Ads. If you need another platform, tell us in the walkthrough form.

Can agencies and multi-brand companies use it?

Yes. The Scale and Custom plans support several stores and brands in one workspace, with each brand's data kept separate.

How long does setup take?

Each connection is a short authorisation. The 24-month historical backfill then runs in the background, and its length depends on your order volume. Forecasts appear once the backfill finishes, provided you have at least 3 months of history.

What are the contract terms?

The trial lasts 14 days. Launch and Scale are billed monthly and can be cancelled at any time. Custom plans may include longer terms by agreement.

What happens to my data if I leave?

You can export your reports as CSV at any time. Ask us by email and we delete your workspace data and confirm in writing.

Walkthrough

See your own cohorts.

A 30-minute call. We look at how your order history fits the model and what the first forecasts would show.

Sent by email, not stored on this site. We reply within two working days.

Ad platforms you use

This opens your email app with the details filled in. See the Privacy Policy.