LTV & retention analytics for Shopify brands

Why is LTV so hard to trust on a Shopify store?

Most Shopify brands have at least three LTV numbers: one from Shopify analytics, one from a retention or email app such as Klaviyo, and one in a founder’s spreadsheet. They rarely agree, because each one makes different choices about revenue versus margin, refunds, discounts, the time window and which customers count.

That disagreement matters because LTV sets how much you can spend to acquire a customer. If the number is inflated by gross revenue and an all-time average, you overspend on channels that bring one-time buyers. If it is too conservative, you underinvest in the channels that bring customers who come back.

The fix is not a better formula. It is a single customer-level dataset built from your own orders, with every assumption written down.

What does a proper LTV and retention analysis include?

The analysis has four layers, and each one answers a different question.

Layer	Question it answers
Cohort retention	What share of customers acquired in a given month are still buying 3, 6 and 12 months later?
Repeat purchase	How long does the second order take, and which first products lead to one?
LTV forecast	What will a customer acquired today be worth in gross margin over 12 or 24 months?
CAC payback	How many months until a customer from each channel covers what it cost to acquire them?

The core calculation is simple: cumulative gross margin per acquired customer, by cohort, over time. A rough version looks like this:

LTV(t) = sum of (net revenue − COGS − shipping − refunds) for the cohort up to month t
         ÷ number of customers in the cohort

The difficulty is in the inputs. Shopify order data includes edited orders, partial refunds, exchanges, draft orders, test orders and customers who checked out under several email addresses. Each of those decisions moves the number, so they are made once, documented, and applied consistently.

How do you forecast LTV for customers who just bought?

Young cohorts have not had time to show their repeat behavior, so a historical average undercounts them. We fit a forecast on the older cohorts, then check it by hiding the most recent months of a cohort you already know and seeing how close the model gets.

Depending on the shape of your data, that can be a straightforward cohort curve extrapolation or a probabilistic customer model of the kind covered in our short history of LTV forecasting. The method is chosen for accuracy on your store and explainability to your team, not novelty.

The forecast is then split by the attributes you can act on: acquisition channel, first product, first-order discount, country, and whether the customer subscribed to email or SMS.

What will I be able to decide afterwards?

The point of the work is better spending and retention decisions. Typical questions the output answers:

1. What is the maximum CAC we can pay on Meta versus Google Ads and still pay back within our target window?
2. Which products are the best entry points for customers who come back?
3. Do first-order discounts bring in customers who retain, or customers who only buy on discount?
4. Where in the customer lifecycle should email and SMS flows push for the second order?

You can sanity-check your own inputs with the LTV calculator and CAC payback calculator before talking to anyone.

Who is this not for?

If your store is less than a year old or has only a few hundred customers, cohort curves will be too noisy to support channel-level decisions. In that case a simple margin-based payback calculation is usually enough, and the tools above will get you most of the way.

It is also not a replacement for an email agency or a retention app. We tell you where retention is won or lost and what a customer is worth; your team or agency runs the campaigns.

When your LTV numbers disagree across tools and you are about to commit real ad budget on the back of them, that is when bringing in Asabisoft for a one-to-two-week diagnostic makes sense.