Why customer retention stalls when ownership is spread across teams, and what a single point of ownership changes.

Retention is shared in its inputs and owned in its outcome. Most organisations only organise for the first half.

In short: Many teams contribute to retention, but the outcome needs one owner. Without one, departmental goals pull against the common number, trade-offs get discovered instead of decided, and retention either drifts or grinds. Businesses already accept this model for product, where a product manager owns the outcome without managing every team that shapes it. Retention rarely gets the equivalent.

Part 1 of this series argued that retention failures usually start inside the organisation, in how well it talks to itself before anything reaches the customer. This part is about the most common reason it doesn’t: retention is treated as everyone’s responsibility, which in practice leaves it with no one.

Consider a leadership meeting during the quarterly review, where retention has decreased by five points based on the definition approved by all stakeholders the previous year. All of them have a slide and a theory ready. Marketing’s campaigns hit their targets, so the drop must have come from somewhere else. Product did its part and engagement held. Finance mentions seasonality and economic environment. Every explanation comes with its own dashboard, and every one may be partly true. The meeting closes with actions for everyone: marketing will test a win-back campaign, product will review the onboarding funnel, finance will model the seasonal and economic effects. Each action follows their own explanation and will be judged on its own team’s dashboard. The following quarter comes and every team can show that its action worked, and… retention is still down.

Retention is shared in its inputs and owned in its outcome

The case for shared responsibility is real. Product decides whether customers find value, support decides what happens when something breaks or someone complains, and marketing, CRM and billing all touch the moment a customer decides whether to stay. That said it seems that treating any single team as “the retention team” is its own mistake.

So, it seems the problem sits at the outcome. Inputs can be shared indefinitely (and sometimes they are). An outcome that belongs to everyone can be claimed by anyone and declined by everyone, usually at the moment it matters (for good or for bad).

This also means that in practice, retention rarely goes entirely unclaimed. It gets claimed by whichever function gravitates closest to it in that particular business model: CRM or marketing in marketing-led businesses, product in product-led ones, customer success in B2B. The others defer, and nobody checks whether the claimant can reach what actually moves the number. Usually it can’t. CRM doesn’t set pricing or design onboarding; product doesn’t run support or billing. The claim covers part of the inputs and none of the rest.

That claim, however, is applied selectively. A good quarter has many parents: the campaign, the release and the new support process all get the credit. A bad quarter belongs to the market, or to another team. Credit flows inward and blame flows outward. Every team’s explanation is locally plausible. Without an owner, however, there is no one forced to reconcile the discrepancy.

Without arbitration, retention drifts or grinds

Where team goals don’t collide, the result is drift. There is a meeting, everyone agrees retention needs attention, nobody in particular is assigned, and a month later the same meeting rediscovers that it still needs attention. An initiative launches with a champion and no owner, and it lives exactly as long as that person’s attention does. The issue is discussed in successive quarterly reviews, phrased slightly differently each time, because whoever names it has no obligation to still be holding it next quarter. When something does work, nobody can say exactly which decision made it work, so it can be hoped for again but not replicated on demand.

Where team goals do collide, the result is friction. Marketing wants more contact, product wants less noise around onboarding, finance wants recovery, risk wants fewer approvals. Each team pulls toward its own target, and the collisions get settled by whoever escalates loudest, or not at all. The common goal is questioned first, then quietly dropped.

The two look different and have the same cause: nobody arbitrates.

Three retention numbers are a symptom, not the disease

Another version of that meeting is one where marketing, product and finance each bring a different retention number, because each counts a different activity: opens, logins, payments. It is the easiest place to see all of this, and the most common place to misread it. The instinct is to treat it as a data problem: agree a definition, fix the dashboards, move on. However that misses what the three numbers actually show. The reason nobody defined retention is because nobody owns it. Each team built its dashboard on its own answer, and the three were allowed to disagree indefinitely because no one was required to reconcile them. It is hard to improve something the business hasn’t agreed exists in a single form.

So the obvious choice is settling the definition which is usually the first thing a real owner does, and so is making every team work from it rather than merely know it. A usable definition answers three questions. First, which activity counts: a login is not a transaction, and an open is not a relationship. Second, over what window, matched to how often this customer naturally returns rather than to the calendar month. Third, whether a customer who only comes back when paid to is retained or rented (to know this is key at start and also in the long run since it prevents a lot of misplaced campaign money). How to measure against the definition honestly is its own subject.

Departmental goals should feed the retention number, not drain it

Departmental goals should exist; a business without them has no way to run its teams. The question is whether each one feeds the common number or draws from it. None of this is new. The team behind Reforge’s growth programmes described the pattern years ago: teams given a metric per stage of the funnel end up optimising at each other’s expense. What’s really missing in most businesses is not awareness of the problem but a way to test for it. The test is simple to ask/ put in practice and uncomfortable to answer: could this team hit its target while retention falls?

Usually, yes:

• A reactivation target measured on returns within seven days can be hit with incentives that bring customers back for a week. By day thirty most are gone again, and the business has rented them rather than retained them.
• An acquisition target on volume or cost per customer can be hit by buying cheaper, lower-intent traffic, whose first-month retention becomes someone else’s number and often impossible to achieve.
• An activation target based on completing onboarding can be hit by making onboarding easy to finish rather than worth finishing.
• A support target on handling time can be hit by closing tickets that aren’t resolved. The customer doesn’t complain twice; they leave.
• A failed-payment recovery target can be hit with dunning aggressive enough to recover this month’s revenue and cause next month’s cancellations.
• A fraud-loss target can be hit by tightening checks until quite a few good customers give up.

No team did anything “wrong” by its own measure and goals. The fix is not to make every team accountable for the whole outcome, which only turns the attribution problem into a formal full on conflict. It is to attach a guardrail to each local target: the reactivation still counts, but only if the customer is still retained by the shared definition at say day sixty. Teams run on fast metrics and are credited on the slower one. Guardrails like these don’t design themselves, though, which is where ownership and authority come back in.

Goals need a hierarchy, and a hierarchy needs someone to enforce it

Individual goals and the common goal can and should coexist (since sometimes one makes us run faster and the other one makes us run longer), and the individual ones matter because they feed the common one. When the two conflict, ranking decides. At the top of that ranking sits the business, not retention. Sometimes retention should lose: to compliance, to margin, or to the honest judgment that a particular customer isn’t worth keeping. That is absolutely fine, provided said ranking is established, communicated and enforced. A ranking that has no one enforcing it is no more than a slide in last year’s strategy deck.

What matters is whether a trade-off is decided or discovered. Retention yields legitimately when leadership (the owner, the executive team or the board, depending on the stakes) decides that it should, with the cost understood in advance. It fails when the decision is made above and the cost is discovered below. Incentives get cut to protect margin, and next quarter the drop in retention is investigated as a CRM problem. Or the trade-off is never decided at all, and each department ends up making its own version of it locally. A trade-off made at the top has to be accounted for at the top.

Retention needs what product already has

Businesses have already solved a version of this problem, for product. A product manager owns the outcome of a product without managing the engineers, designers or analysts who build it, and nobody finds that arrangement strange. Bear in mind this analogy has a limit worth stating: a product manager’s leverage comes from setting the priorities of a team dedicated to that product. A retention owner (as things stand in most organisations) has no such team, so the leverage has to come from a mandate given by the executive team over three things: the definition of retention, the ranking of goals, and the guardrails each team works within. Retention, which in a high-frequency business sits about as close to revenue as anything does, rarely gets either the role or the mandate.

The misconception for many businesses is they believe they already have someone that solves that problem because they have a Head of CRM or a lifecycle lead. Although usually that role is scoped to channels: the messages, the automation, the send/ promotional calendar. It owns what goes out, not whether customers stay.

Other businesses do have a Head of Retention, and then it depends on what the title can reach. If the role has real cross-functional authority, much of this piece may not apply, and that’s worth saying plainly. If it’s CRM under a new name, still scoped to channels, the problem is unchanged under the hood of an euphemism that solves nothing on its own. If it has the remit on paper but not the authority in practice (more often than not the case), that’s a different failure, covered below. The test is simple: can the role change anything outside its own team?

The equivalent of a product manager for retention does a few specific things:

• It holds the definition of a retained customer and reconciles the dashboards against it.
• It designs and maintains the guardrails on each team’s targets: which ones apply, at what threshold, and what happens when two of them pull against each other. This is key and is most of the job, and it is how departmental goals get ranked against the common number instead of draining it.
• It arbitrates contact with the customer. CRM’s frequency rules govern CRM’s channels; when product sends from its own tool and billing sends from another, someone has to put every sender under one contact policy. When nobody does, the customer arbitrates for you, usually by muting you.
• It gives internal communication somewhere to land. Part 1 was about teams that don’t talk to each other; here the teams talk perfectly well, and what they say circulates without converging unless someone receives it, reconciles it and turns it into a decision.
• It explains, on a schedule, why the number moved. When the cause sits beyond any team’s campaigns, such as a product losing its value, the owner can’t fix it alone, but is the one obliged to say so early and with evidence.

None of this requires a new department, and often not new headcount; more often it means one existing person given the job the number already implies. Nor does it mean routing every decision through that person. The owner sets the rules teams decide within, and steps in only when those rules collide.

The parallel has a known weak point, though. The most common complaint from product managers is responsibility without authority, and a retention owner faces the same risk: their name attached to the number, without the standing to change what moves it. That is Part 3.

How to tell whether retention in your business is actually owned

A few questions answer it faster than an org chart:

• Could three of your leaders give the same retention number, calculated the same way, unprompted?
• If retention moved this quarter, is there one person whose job it is to explain why, or does the explanation arrive in pieces from separate reviews?
• Could every team hit its targets while retention falls?
• When two teams want to contact the same customer on the same day, who decides?
• The last time retention was traded off against something else, who made the call, and did anyone account for the cost?

Uncomfortable answers are not always about ownership. Three unreconciled dashboards can also mean the data genuinely won’t reconcile: if no system holds a single view of the customer, even a capable owner has nothing to reconcile against, and that is a tool constraint. The quickest way to tell is to give one person the job of reconciling for a month. If they can’t because the data won’t join, the binding constraint is the stack. If they can and nothing changes, it’s ownership.

The point

Retention is not a department, and it can’t be run as a shared one. It is a number, and a number needs one name attached to it, or accountability dissolves into something that looks like consensus and behaves nothing like it.

Contribution is everyone’s. The outcome is owned in one place. Arbitration is what keeps both of those true. None of this is a new idea, since most businesses already run their products this way. If asking who owns retention in yours gets you a discussion rather than a name, that’s worth talking through.

Frequently asked

Isn’t retention supposed to be everyone’s responsibility? Everyone contributes to it, which is different from everyone being accountable for it. Product, support, marketing and CRM all shape whether customers stay, but the outcome still needs one owner who defines it, reconciles the teams’ views of it and answers for its movement.

What counts as a retained customer? There is no universal definition, which is exactly why each business needs one that every team uses. A usable definition specifies which activity counts, over what window (matched to the customer’s natural cycle), and whether customers who only return when incentivised count as retained.

Isn’t the Head of CRM the retention owner? Usually not. The role is typically scoped to channels: messages, automation, the send calendar. It owns what goes out, not whether customers stay. Owning retention means owning the outcome across the teams that shape it, the way a product manager owns a product.

We have a Head of Retention. Does this still apply? It depends on what the role can reach. With real cross-functional authority, the problem described here may not exist. If the role is CRM under a new title, or has the remit without the authority, it does. The test is whether the role can change anything outside its own team.

Wouldn’t guardrails on each team’s targets be enough? Guardrails are the main mechanism, but they don’t maintain themselves. Someone has to hold the definition they rely on, set the thresholds, notice when two teams’ guardrails contradict each other, and adjust them as the business changes. That is the owner’s job; without it, the guardrails drift the same way retention does.

How is this different from Part 1? Part 1 is about teams and systems that don’t communicate, so errors reach the customer unseen. This part is a level up: even when teams communicate well, if nothing converges on a single owner of the outcome, the pieces stay coordinated locally and the number still drifts.