Every few weeks a founder asks us some version of the same thing: "Klaviyo keeps showing me this Personalization add-on. Do we need it?"
My answer is usually yes, but with a condition. It's worth paying for if you're going to use RFM to trigger a flow in the first month. If the plan is to open the report now and then and look at the charts, skip it. You'll pay for a dashboard nobody looks at after the second week.
Here's what it is, what it costs, and how we actually use it.
What Personalization is
Personalization is the add-on Klaviyo used to call Marketing Analytics. It was renamed around K:BOS in September, and a lot of the help center still uses the old name, so if you go looking for setup docs, search for both. Pricing starts at $100 a month and goes up with your active profile count, so check the slider for your size before you decide.
It's worth knowing what you already get without it. Standard Klaviyo gives you predictive analytics on each profile: predicted CLV, churn risk, average time between orders. You need at least 500 customers who've ordered, 180 days of order history, and orders in the last 30 days for those to show up. That covers a lot of "who's about to drift" questions already.
What the add-on adds on top:
- RFM analysis. Every customer gets put in one of six groups based on how recently they bought, how often, and how much they spend.
- Customizable CLV. You can change the prediction window instead of taking the default.
- Attribution model choice. By default Klaviyo uses last-touch attribution. With the add-on you can change the model.
- Multiple RFM models. As of September you can run up to four separate RFM models, for example one for subscription orders and one for one-time orders. A customer can be a Champion in one and At risk in the other.
RFM is the part that pays for the rest. So that's what I'd judge it on.
How the RFM groups work
Klaviyo scores every customer 1 to 3 on recency, frequency and monetary value. By default, recency is 3 if they bought in the last 180 days, 2 if they bought in the last 365, and 1 if it's been longer than that. Frequency and monetary are split into thirds of your own customer base. The three scores together put each customer in a group:
The help center has suggested actions for each group, and they're fine. But the groups only matter if something happens when a customer moves from one to another. That's the part most accounts skip.
What we actually see in accounts
In the Klaviyo accounts we work in, segments built on RFM groups are pretty common. A live flow that fires when someone's group changes is rare.
So in most of those accounts, someone set up RFM, made the six segments, maybe used them to filter a campaign or two, and stopped there. RFM ended up as a report and a set of segments. Nothing fires when a good customer starts to drift.
Turning RFM into a trigger
Every profile gets three RFM properties: Current RFM group, Previous RFM group, and RFM group last changed. They refresh every night. That's all you need to set it up.
- Build a segment where Current RFM group equals Needs attention OR At risk. Add "is subscribed to email" if you want a clean send list.
- Start a flow triggered by that segment. Segment-triggered flows fire when someone newly qualifies, so a customer goes in on the night they slip into one of those groups, not on some arbitrary day.
- Write it for who they are. Needs attention customers spent a lot or bought often and then stopped. Lead with what's new or what they bought last, and test whether you need a discount at all before you add one. At risk customers spent less, so Klaviyo's own guidance is to show them cheaper products and not over-message.
The same trick works going up the ladder too. A segment for "became Loyal or Champion" is a good place for a review request, because those are the customers who already like you enough to say so.
What it did in one account
In one account where this is wired up, we built two flows off RFM in December 2025:
- A churn-risk flow for customers who move into Needs attention or At risk.
- A review request for customers who become Loyal or Champion.
Over the last year the churn-risk flow sent to about 54,600 recipients, drove 127 orders and $15,376 in revenue Klaviyo attributes to it. That works out to $0.28 per recipient.
I want to be straight about how that compares. The same brand also runs a standard customer winback, the kind that's based on time since last order. Same period, it made $1.40 per recipient and about $81,000.
So the RFM flow didn't beat the basic winback, and it didn't replace it. What it did was add a second catch at a different moment. The winback fires on a timer. The RFM flow fires when Klaviyo sees someone's buying pattern change, which for a frequent buyer can happen well before the timer would. That's about $15K in attributed revenue on top of the winback, from a flow that runs itself.
Whether that's worth it for you comes down to simple math. Find your price on the slider, multiply by 12, and compare it to what one more flow at a few cents to a few dimes per recipient would bring in across your churned customers. Every brand's customers are different, so your per-recipient number could land higher or lower. A supplement brand with repeat buyers is a good fit for RFM. If most of your customers buy once, frequency scores don't tell you much and the groups get blurry.
When I'd skip it
- You don't meet the data requirements. RFM needs at least 500 customers who've placed an order, 180 days of history, recent orders, and some customers with 3 or more orders. Below that, the groups aren't meaningful.
- Your catalog is basically one-and-done. Furniture, big equipment, a single hero product people buy once. Predicted CLV and a timer-based winback will do most of the work for free.
- Nobody is going to build the flow. If it sits in a backlog for three months, you've paid $300 or more for six segments.
What I'd do this week
If you already pay for Personalization, go to your segments list and search "RFM." Then check whether any live flow is triggered by one of those segments. If none are, that's the first thing to fix, and it's maybe an afternoon of work.
If you don't pay for it yet, start with the free predictive properties and a winback on a timer. Once that's running and earning, RFM is the next layer.
Which brings me to the question I'd ask any team thinking about it: if you turned it on tomorrow, which flow would you build first?
