LTV Measurement
The methodology hub for LTV measurement in online retail: which formula to use, when cohorts beat averages, and how to avoid the contribution-margin trap.
LTV Measurement
LTV measurement is the methodology used to estimate the total contribution a customer generates over their lifetime with your store.
LTV measurement is the set of choices you make before you compute a single number: which time horizon to use, whether to forecast forward or look back, whether to measure top-line revenue or contribution margin, and whether to bucket customers into acquisition cohorts or smear them across a sitewide average. Get those choices right and LTV becomes the most reliable input into your CAC ceiling, paid-media bidding, and merchandising decisions. Get them wrong and you end up overpaying for traffic that looks profitable on paper and bleeds cash in reality.
Most stores compute LTV exactly once — usually in a board deck — and then never touch the methodology again. That's a problem, because the four upstream choices below quietly change the answer by 2-3x.
This page is the methodology hub. It covers the three decisions that matter most, links to the deeper resources for each (Predictive LTV, the LTV Calculator, LTV Benchmarks by Industry), and gives you a defensible default for each choice so you can start with something sensible and refine later.
Predictive vs historical LTV
Historical LTV looks backwards: sum the revenue (or margin) every customer has generated so far, divide by the number of customers, and you're done. It's accurate for the cohorts that have fully matured, but useless for customers acquired in the last 90 days — they haven't had time to repeat yet.
Predictive LTV projects forward from observed early behaviour (first-order AOV, time to second purchase, category bought) and estimates what each customer will be worth at 12, 24, or 36 months. It's what you need for live paid-media bidding, because Meta and Google can't wait three years to learn whether a buyer was profitable. Use historical to calibrate, predictive to bid.
Revenue LTV vs contribution-margin LTV
Revenue LTV is what most Shopify dashboards show you — total spend per customer over X months. It's a flattering number and almost always the wrong one to compare against CAC. A customer worth €240 in revenue at a 35% gross margin and €18 of fulfilment per order is worth roughly €60 in contribution, not €240.
Contribution-margin LTV nets out COGS, payment fees, fulfilment, returns, and any per-order variable cost. That's the number that should sit above CAC in your unit-economics model. If you're running paid acquisition against a revenue LTV figure, you're effectively bidding with someone else's money — until the cash runs out.
The contribution-margin trap
An apparel store we audited had a stated LTV:CAC of 4.2 and was scaling spend aggressively. Once returns (28% on dresses), fulfilment, and payment fees were stripped out, the real ratio was 1.4 — below the 3.0 threshold most operators target. They weren't winning; they were borrowing growth from working capital.
Cohort LTV vs sitewide average
A sitewide LTV average hides everything that matters. The customer who came in through a Black Friday 40%-off promo behaves nothing like the customer who came in through organic search on a full-price hero SKU. Average them together and you get a number that describes neither, and that overstates the value of your worst-performing channels.
Cohort LTV groups customers by acquisition month (or channel, or first-product) and tracks each cohort's retention curve separately. You then have a defensible LTV per channel, which is what you need to set channel-level CAC targets. The LTV Benchmarks by Industry page shows what mature curves look like in beauty, apparel, supplements, and home goods — use those as sanity checks against your own cohorts.
Cumulative contribution per customer by acquisition channel (24 months)
Organic / direct
Paid social (full-price)
Discount / promo cohort
LTV measurement: common questions
Average order value × purchase frequency × gross margin × expected customer lifespan in years. It's a historical, contribution-margin-aware estimate that's good enough for a first pass. The LTV Calculator walks through the inputs and gives you a number in under a minute.
Both, for different jobs. Historical LTV is your ground truth for cohorts older than 12-18 months and your benchmark for accuracy. Predictive LTV is what you bid with on Meta and Google, because you can't wait two years to learn whether a click was worth it. Calibrate the predictive model against historical cohorts quarterly.
It depends on repurchase cycle. For consumables (supplements, coffee, skincare refills) 6-9 months captures most of the curve. For apparel and home goods, you need 18-24 months. For considered-purchase categories like furniture, 36 months is closer to the truth.
Yes, always — net out return-related refunds, restocking costs, and reverse-logistics fees at the order level before you compute contribution. In categories like apparel and footwear, returns can be 20-35% of gross revenue, and an LTV figure that ignores them is the single most common reason stores overpay for paid traffic.
3:1 is the common rule of thumb for sustainable growth, with 1:1 being break-even on a fully-loaded basis and anything above 5:1 usually meaning you're under-investing in acquisition. Make sure both sides use contribution margin, not revenue, or the ratio is meaningless.
Subscription LTV is largely a churn calculation: ARPU divided by monthly churn rate gives you the lifetime estimate directly. One-time-purchase LTV needs a repeat-purchase model (frequency × lifespan) because there's no recurring billing signal. Mixed-model stores should compute the two segments separately and weight them by acquisition mix.
Treat it as a ceiling, not a target. Shopify's LTV is revenue-based, sitewide-averaged, and historical — three of the four choices we'd push back on. It's fine for trend-watching, but don't bid against it without first converting to a cohort-level contribution-margin number.
Marketplace customers (Amazon, Zalando) are largely invisible — you don't own the email, the repurchase signal, or the channel attribution, so LTV defaults to first-order contribution minus marketplace fees. Direct-to-consumer sales let you measure the full curve. Most operators run two parallel LTV models and don't average them together.
Discount-acquired cohorts repeat less, refund more, and have lower full-price tolerance for at least 12 months. If you average them with full-price cohorts, your sitewide LTV is inflated by the volume and depressed by the quality. Always segment by first-order discount band.
The methodology itself you revisit yearly. The numbers you refresh monthly for predictive LTV (so paid-media bidding stays calibrated) and quarterly for historical cohort curves. A static LTV figure in a deck from 18 months ago is worse than no LTV figure at all, because someone is probably still bidding against it.
Start tracking the metric that matters
Free for 30 days. No credit card. Connect your site in 4 minutes and see the one metric driving revenue.