Kris is the co-founder of Mage Loyalty. I spend most days talking to merchants, shipping features, and making sure our customers get real results. If you run a Shopify store or Agency we should chat!
The economics
Why customer value matters more than acquisition cost
It is easy to focus on return on ad spend, email revenue or conversion rate. But once you have paid to acquire a customer, only one question is left: how many more purchases can you earn from that customer?
Here is a simple example.
Both brands pay exactly the same to acquire a customer. Brand B earns $60 more from each one in the first 90 days.
That gap matters because customer value sets how much you can afford to spend on acquisition. Brand B can bid more aggressively on ads and scale further. Brand A is capped by the smaller amount it earns back from each customer, whether you measure over 90 days, 180 days or any other window.
Find the constraint before you pick a tactic
When brands want to become Brand B, they usually reach for tactics first: launch a loyalty program, send more emails, run more Klaviyo campaigns, add win-back flows.
These are all good strategies, and they can all work. But sending more emails is not automatically better retention, and blasting too many can push customers away. Before you implement anything, work out where the constraint on your 90-day gross profit is. That starts with the metrics.
The metrics
The five retention metrics and how to calculate them
Average order value
AOV is the average basket size across all your orders, or across the orders in a set time period. For example, $500,000 in revenue across 10,000 orders gives you a $50 AOV.
On its own, AOV tells you nothing about margin. Always read it alongside your average gross profit per order, so you know what you actually keep.
Repeat purchase rate
RPR is the percentage of customers who bought once and then came back to buy a second time. 10,000 first-time customers, 3,000 of whom buy again, gives you a 30% RPR. If 5,000 came back, it would be 50%.
Always fix a window rather than measuring across all time. The best window is 90 days, because it lines up with quarters. You can compare this quarter with last quarter, and with the same quarter last year, and see clearly whether you are trending up or down. Use 12 months as the maximum window.
Time to second purchase
This measures how long it takes a customer who has bought once to come back and buy again. Always use the median, not the average. Line up every customer's gap between first and second order and take the middle value. Don't add all the gaps together and divide by the number of customers.
Some customers will always come back after a very long time, and they skew the average badly. The median tells you when customers are really returning.
Purchase frequency
The number of purchases per customer over a fixed period. Measure it over 12 months for the most accurate picture.
90-day customer value
What a customer is worth to you in the 90 days after you acquire them. Ideally, measure this as gross profit rather than revenue, because that is what you actually keep. If you can't get to gross profit yet, start with revenue and refine it later.
Time to second purchase: one of the most underrated metrics
A lot of brands look at repeat purchase rate on its own and stop there. Take two brands, both with a 35% RPR. On paper they look identical. In practice they are completely different businesses.
Brand A recovers its acquisition cost much faster, generates cash sooner to reinvest into acquiring the next customer, and has room for a third purchase in the same window Brand B spends waiting for a second.
Small changes compound
Now look at purchase frequency across a year, with 10,000 customers and the same $50 AOV at both brands. Brand A averages 2.3 orders per customer a year, Brand B 2.0. A difference of 0.3 orders per customer looks tiny.
| 10,000 customers · $50 AOV | Brand A | Brand B |
|---|---|---|
| Orders per customer per year | 2.3 | 2.0 |
| Total orders | 23,000 | 20,000 |
| Revenue | $1,150,000 | $1,000,000 |
That 0.3 difference adds up to 3,000 extra orders and $150,000 in extra revenue for Brand A.
On paper, with identical RPR and identical AOV, you would assume these brands perform the same. That's why you need to look beyond AOV, conversion rate, CAC and RPR to the full set of metrics.
How the metrics fit together
Think of it as a tree. Customer value sits at the top. Below it sit purchase frequency, repeat purchase rate and average order value. Below those sits time to second order.
Customer value, purchase frequency, RPR and AOV are behaviours: how your customers actually buy. Time to second order and 90-day customer value are economic values.
Moving any of these in the right direction means your retention is improving. Open rates, sign-ups and first purchases don't tell you that on their own. What matters is how much revenue you earn from each customer, and how quickly.
What drives retention
The retention hierarchy
The metrics show where you stand. Other factors decide whether your retention performs. At Mage, we look at them as the retention hierarchy: the five factors that decide whether a customer comes back, in order of importance.
- Level 1ProductIs the product good enough to buy again?
If the product isn’t good, the customer won’t come back, however strong your retention strategy is. No win-back campaign fixes a poor product.
- Level 2PortfolioIs there anything else for the customer to buy?
If a customer has bought from you and there is nothing left to sell them, they have no reason to return. Your range, collections and upsells decide whether there is a next purchase to make.
- Level 3Acquisition qualityAre you acquiring customers who will stay?
Not every customer is the same. A customer who buys at full price for $150 will have a very different lifetime value from one who only buys at $40 during Black Friday or a seasonal sale. Know who you are acquiring.
- Level 4Customer experienceIs buying from you a good experience?
Delivery, service and support, and whether you replace a faulty product. You can get levels 1 to 3 right, but a poor experience still leads to bad reviews, unhappy customers and a damaged reputation.
- Level 5Amplifies 1 to 4Retention systemAre email, SMS and loyalty bringing customers back?
This is the level most brands treat as the only one that matters, and it is the least important. You can have the best loyalty program and the best win-back flows in the world, but they won’t work until levels 1 to 4 are right.
Retention marketing doesn't create value on its own. It amplifies the value you have already created everywhere else in the brand.
The biggest constraint: the first to second order
Many brands acquire a customer, and then the customer either doesn't come back at all or doesn't come back fast enough. Treat both problems as equally important. A customer who first buys in January and returns 365 days later isn't earning back your acquisition cost fast enough to reinvest and scale.
A typical customer journey looks like this:
- Visitor
- First orderThe gapThe hardest step: 1st → 2nd order
- Second order
- Third order
- Loyal customer
- StepVisitor
- OrderFirst order
- The gap
- OrderSecond order
- OrderThird order
- ResultLoyal customer
The most important step is turning the first order into a second. Once a customer has placed a second order, the relationship changes:
- Trust: they trust your brand.
- Familiarity: they know your products and your range.
- Data: you probably have their email and SMS, so you can remarket to them.
Getting a customer from their second order to their third is much easier than getting them from their first to their second.
Three questions to ask about your first-to-second gap
- What percentage of customers go on to place a second order?
- What is the median time between the first and second order?
- Where are the rest dropping off, and why?
Common reasons for the drop-off:
- They are going to a competitor.
- Your email marketing isn’t strong enough.
- You aren’t capturing enough data to remarket to them.
- You haven’t built enough trust.
- Your reviews aren’t prominent enough, or you need more UGC.
- There isn’t enough incentive to come back.
Once you know where customers drop off, you can work out why.
The levers
Lever 1: increase AOV, profitably
There are three ways to grow customer value. Customers can spend more, buy more often, or keep buying for longer.
The best strategies work on all three at once. A common mistake is to invest heavily in an agency for great Klaviyo flows, and then forget incentives and purchase frequency altogether.
Increasing retention also increases AOV. By the second or third purchase, customers trust you and know your range, so they are more comfortable adding more to their basket, provided your portfolio gives them something to add.
Six ways to move AOV
- Product bundles
- Cart goals: a progress bar in the cart, such as “spend X, get free shipping” or “spend X, get a free product”
- Cross-sells: spot product pairings and show them to the customer
- Loyalty incentives: store credit back on a purchase, or a stamp card (buy X, get Y). These lift RPR, which lifts AOV indirectly.
- Free shipping threshold: works particularly well
- Free gift
You can run one of these or all of them, depending on your goals.
Make sure the AOV is profitable
Don't lift AOV just to move the metric. You can't offer cart goals, free shipping and loyalty incentives if your margin doesn't allow it.
Gross profit per order = AOV × margin
If you raise AOV from $50 to $60, the question is how much of that extra $10 is profit: $5? $6? That is the number that actually affects your business.
Lever 2: bring customers back sooner
The goal isn't only to get customers to come back. It's to get them to come back sooner, because shrinking the time between orders is one of the strongest drivers of compounding revenue. Sooner comes first. Coming back at all comes second.
What not to do
The classic mistake: a customer buys, 15 to 30 days pass without a second purchase, and a Klaviyo flow sends them a discount. The aim is to close that gap without reaching for a discount.
What to do instead
Reward customers more as their lifetime value grows, usually through VIP tiers. The rewards don't have to be discounts. Scaled percentage discounts can work in some models, but the rewards performing best right now are exclusive events, early access to sales and launches (including early Black Friday access), and above all exclusive VIP products.
The aim is an exclusive club the customer feels part of, with a spend or order target to hit to keep their status. A rolling 12-month window works particularly well, rather than a calendar year or lifetime status. Orders older than 12 months (or 6 months, if you prefer) drop off, so the customer must keep spending to keep their tier. That gives them a reason to keep making purchases.
If you run subscriptions, look at shortening the default cycle: for example, from 60 or 90 days down to 40, 45 or even 30 days. Some categories can support a 14-day cycle. It depends on your product.
Market to customers based on what they first bought. If their first purchase came from one collection, show them a relevant category they haven't tried yet, instead of a generic campaign covering the whole store. Relevance shortens the time to a second purchase. Take it further by segmenting on VIP tier, so each tier gets marketing that suits its benefits. This works for the fifth, tenth and twentieth purchase too, not just the second.
Set a reward or tier threshold, based on your expected AOV, that the second order always reaches. Then you can tell customers they are one order away from a free gift, a free product, an exclusive discount or an event invite. A progress bar in the email that shows what they will unlock works well, and a free product performs particularly well. Structure this correctly and nearly every second order unlocks something.
Why this beats a discount
You may still discount later in the journey. The difference is that you don't give the discount away up front to win the second purchase. You secure the second purchase first, then reward it.
By then you have earned the customer's trust, they know your products and your range, and you have their data. That makes closing the gap from second order to third far easier. Train customers that the more loyal they are, the more they are rewarded, not that waiting brings a discount.
Discounts vs incentives
The old strategy was to offer as many discounts as possible. The new strategy is to offer as many incentives as possible.
The discount cycle
This is one of the biggest mistakes in ecommerce. RPR is low, so the brand sends 20% off, then 15% off, then 20% off again, and the cycle repeats.
Customers learn to wait for the next code. If the discounts grow over time, they learn that the longer they wait, the better the offer. That lengthens your time to second order, which is the opposite of what you want.
Reframe the question
Instead of asking “what discount can we offer to get a purchase?”, ask “what value can we exchange for another purchase? What reason can we give the customer to come back?”
Incentives that work
| Incentive | Cost to margin |
|---|---|
| Early access to launches and sales | None |
| Exclusive VIP products | Depends on product costs |
| Perks and benefits, e.g. free engraving or gift wrapping for VIPs | Low (depends on fulfilment cost) |
| Free shipping | Some |
| Automatic shipping upgrade at checkout for top VIP tiers | Some |
The automatic shipping upgrade is performing especially well right now. If your shipping costs are low, upgrading shipping is often cheaper than giving 10% off. Check your margins first, because it isn't always possible if your shipping costs are high.
You could argue the margin-cost options are still a form of discount. But they usually cost less than a constant discount cycle, and there is a key difference: every one of these is earned. Customers unlock them by hitting a spend target, an order count or a points balance. They aren't handed out to everyone.
The faster customers come back, the faster they unlock more benefits. You stop cutting your price and start building perceived value.
Loyalty as a system
Loyalty as a behavioural system
A loyalty program doesn't give you a new metric to track. It moves the metrics that already matter.
The last row is the one that drives revenue. You aren't only raising RPR, frequency and AOV; you are raising how quickly they happen.
Points alone aren't a strategy
Giving customers points and hoping they come back is a discount program, not a strategy. A loyalty program should be designed to change a specific customer behaviour:
BehaviourIncentiveActionValue
Start from your constraint and design the program around it:
That last row is easy to overlook. Even if customers already come back and your AOV is healthy, you still need loyalty. Your highest-value customers champion your brand, refer their friends and are a key revenue driver. You don't need to discount to keep them, but they do need to feel they are earning something for being loyal.
Loyalty isn't a discounting system. It is a behavioural system: understanding how customers interact with your brand and influencing that positively. That is the difference between a winning strategy and a points program that does little for your brand and probably hurts your margins.
Protect the relationship with your best customers
From the customer's side, the ideal journey runs from a new customer activating their account, through the first and second order, to more frequent orders, a repeat buyer and a VIP. By the end of that journey, the customer trusts your brand, likes your products and your range, and may be referring friends and posting about you. At that point, your priority is to protect the relationship.
That means the experience has to change as the customer progresses. A VIP placing their tenth order should not have the same buying experience as a first-time customer. If they do, they won't feel valued. You will struggle to keep them, and you'll be stuck acquiring new customers to replace them, which costs more than keeping the ones you have.
This is where generic loyalty programs fall down: every customer gets the same experience. A first-time customer earns one point per dollar, and a VIP on their tenth order earns one point per dollar too. That does nothing to keep your VIPs for the long term.
Instead, scale rewards or access as customers progress, with some form of exclusivity. That links back to the incentives above: early access and exclusive products. If you use discounts, offer enhanced discounts to your most valuable customers. It may not help your margin up front, but in the long run it protects your margin far better than losing those customers.
The question to ask isn't “how do I extract more revenue from a VIP?” It's “how do I retain this VIP?” If a customer is already buying constantly, you don't need to worry about extracting more value. You need to protect the relationship.
Track tier progression, not enrolment
Here is a metric many brands don't track: the percentage of customers who move up your tiers.
That isn't the percentage enrolled in your loyalty program. Membership is only the first step. It is useful for collecting first-party data and remarketing, but the real value comes from behaviour change: customers actually progressing through your tiered rewards or perks.
Use your VIPs to find more VIPs
Once you have data on the customers in your VIP tiers, you know exactly what a high-value customer looks like: who reaches your spend thresholds, and which customer types reach your top tier and produce the most value. Take that back to acquisition and ask:
- What product did they buy first?
- Which products along the way led them to buy again?
- Which channel did you acquire them from? Customers from one channel may reach higher tiers than customers from another.
You no longer have just conversion data on your acquisition. You have a whole lifetime of data on those conversions, which tells you where to invest to acquire more high-value customers.
The data loop
Many brands treat their loyalty program and subscriptions as separate from their ads and acquisition, as if existing customers have nothing to do with the ones not yet acquired. In fact, it is all one connected data loop, and you can use it to see what is performing and what isn't.
- Behaviour: whether customers buy, don't buy, or perform really well with your brand.
- Data: that behaviour gives you data. Every interaction with your loyalty program is data.
- Segments: split customers into groups based on how they behave.
- Messaging: send each segment different messaging.
- New behaviour: customers respond to that messaging, and the cycle repeats.
This data is probably the most valuable part of your loyalty program. It shows you:
- Who is increasing their spend
- Who has stopped spending, dropped out of the top tier or stopped buying altogether
- Who responds well to rewards
- Who refers their friends
- Who is moving towards your top tier
It also feeds into everything else you run: your ad campaigns, your subscription campaigns and your wider brand strategy.
Better Klaviyo segments with loyalty data
Most brands use Klaviyo on a surface level: standard campaigns, standard win-back flows, offers based on purchase history. But what you get out of Klaviyo depends directly on the data you put in.
That lets you split your messaging by what each customer responds to. One segment gets reward-based incentives, while another that responds better to perks or exclusive access gets those instead.
Loyalty data expands your segments sideways, giving you new ways to group customers. For example:
- Customers who may be going inactive
- Customers with lots of unspent points in their account
- Customers close to reaching the next tier
- Customers about to drop down a tier
- Customers who respond best to rewards, and those who respond best to perks or exclusive access
Customers who behave differently shouldn't all get the same messaging. Segmenting them by behaviour, so each group gets messaging relevant to them, is only possible with the data loop in place.
