Retention is a marketing job too: fit before the sale, visible value after it

1. Blog
2. Conversion & Pricing

By Saroj Jha · October 29, 2026 · 12 min read

Most startups treat marketing as the team that wins customers and hands them over. After the contract is signed, customer success, support or account management take the relationship, and marketing goes back to filling the pipeline. In my view that split leaves the cheapest revenue a startup has, the revenue it already won, without anyone shaping what customers learn, believe and repeat after the sale.

Marketing changes retention in four places: who it brings in, how quickly new customers reach a first result, whether the person who holds the budget can see that result, and how expansion is offered. None of this means marketing takes over customer success. It means marketing designs what customers are told and shown after the sale, using the same skills it uses to win them.

Running case: Example Co. Fictional

Example Co. is a seed-stage B2B SaaS company that sells an invoice-approval workflow to finance teams at companies with 50 to 500 employees. It started the year with 160 customers and $1,000,000 in annual recurring revenue (ARR), an average of $6,250 per customer. Marketing runs paid search, webinars and a newsletter, and hands every signed customer to a two-person customer success team. Example Co. and every number attached to it are illustrative.

Why should marketing care what happens after the sale?

Because the revenue a company keeps is the base everything else is added to. Net revenue retention (NRR) measures what the customers you had a year ago pay you now, including expansion; gross revenue retention (GRR) measures the same thing without expansion. In SaaS Capital’s 2025 survey of more than 1,000 private B2B SaaS companies with at least $1M ARR, the median NRR was 101% and the median GRR was 91%, measured from December 2023 to December 2024^[1]. Its April 2026 update for bootstrapped companies with $3M to $20M ARR puts the medians at 103% and 91%, essentially flat on the year before^[2].

Running case: Example Co. Fictional

Over the year, 20 of Example Co.’s 160 customers cancelled, taking $110,000 of ARR, and downgrades removed another $40,000. Expansion from 24 accounts added $200,000. NRR is ($1,000,000 − $150,000 + $200,000) ÷ $1,000,000 = 105%, which looks healthy. GRR is ($1,000,000 − $150,000) ÷ $1,000,000 = 85%, six points below the 91% median, or $60,000 a year more lost than a median company on the same base.

A 105% NRR can hide a base that leaks, and the expansion that covers the leak comes from a few accounts. I go through why in Your NRR is 100% and you are losing a quarter of your customers, so here I only take the point that matters for marketing: read GRR first, because it shows what the base keeps on its own.

You will also see the claim that 5% more retention lifts profit by 25% to 95%. The version most often quoted is from a 2000 Harvard Business Review article by Frederick Reichheld and Phil Schefter about online companies: “By retaining a mere 5% more customers, e-companies can boost profits by 25–95%”^[3]. A later Bain article by Reichheld and Christine Detrick is narrower: “In financial services, for example, a 5% increase in customer retention produces more than a 25% increase in profit”^[4]. Neither studied B2B SaaS startups, so I read them as a direction, not a forecast for your company.

Do this

Compute GRR beside NRR for the last twelve months before you plan any retention campaign. The Churn and Retention Calculator takes starting revenue, expansion, contraction and churn for one period.

Free tool

Put GRR beside NRR on your own numbers

The free Churn and Retention Calculator turns one period of starting revenue, expansion, contraction and churn into NRR, GRR, logo churn and the quick ratio.

Which churn can marketing actually influence?

Only the churn whose cause sits in what customers were promised, taught or shown. More emails will not fix a missing capability, and a better onboarding sequence will not save a customer who never had the problem you solve. So I start with the cause of each loss, then decide what marketing contributes and who it has to work with.

The point of the table is to force one driver per lost customer. The team then argues about evidence instead of opinions, and each driver has one owner and one partner.

Running case: Example Co. Fictional

Example Co.’s founder called or emailed all 20 customers who cancelled and tagged each with one driver. Eight were a poor fit ($36,000 of ARR), six never activated ($30,000), four could not show value at renewal ($32,000) and two lost budget in a reorganisation ($12,000). That is 20 customers and $110,000. Fourteen of the 20 left for the two reasons that start in marketing’s own work: who was sold to and what they were taught first.

Do this

Tag every customer who left in the last two quarters with one driver from the table. Use the Customer Interview Kit for the calls, and cancelled customers’ own words, not the reason picked in a cancellation form.

Retention starts with who you sell to

If marketing brings in customers the company cannot reliably make successful, a good acquisition quarter becomes a bad renewal quarter. That is why I define the ideal customer profile (ICP) by who succeeds, not only by who buys. Compare your best-retained customers with the ones who left, and look for the traits the first group shares.

Each trait then needs a check that marketing and sales can apply before the contract: a qualifying question on a demo form, a line in the discovery call, a use case on the pricing page that tells poor-fit buyers this is not for them. The ICP and Account-Scoring Playbook turns these traits into weighted scores, and the ICP Fit Scorer tiers accounts against them.

Running case: Example Co. Fictional

Of Example Co.’s eight poor-fit losses, five bought for a one-off audit clean-up and three had no named sponsor in finance leadership. Marketing added two qualifying questions to the demo form (“How many invoices do you approve each month?” and “Who owns the approval policy?”) and rewrote the pricing page to say the product is built for recurring approval volume.

Do this

List your 20 best-retained customers and the customers who left last year, and mark which of the four traits each had at the time of purchase. Keep a trait in your ICP only if it separates the two lists.

How does education get customers to first value?

Education for retention is not more documentation. It moves each customer to a specific first result, the moment the product has visibly done its job for them, and it reaches each role with what that role needs. The end user needs workflow help; the executive sponsor needs evidence that the purchase is working.

• Define activation. Name one observable milestone that proves value, such as the first campaign published or the first transaction processed, and a time window for reaching it.
• Write by role. Send the end user how-to guidance and the sponsor a short progress note, not the same email to both.
• Trigger on behaviour. Send guidance when an account stalls before the milestone, not on a fixed day-3, day-7 calendar.

Running case: Example Co. Fictional

Example Co. defined activation as the first approval policy live and 25 invoices approved through it within 30 days of signing. Accounts that had not set a policy by day 10 got a short setup guide for the finance operations lead and a one-paragraph note to the sponsor naming the step still open.

Do this

Write down your activation milestone and the window for reaching it, then check which of your lifecycle emails move an account toward it. The Lifecycle Email Audit lists every send against the stage it serves, and the Customer Onboarding and Retention Playbook covers the onboarding steps.

How do you make value visible to the person who renews?

Customers can get value and still be unable to explain it to the person who holds the budget. Usage numbers such as logins and reports generated are not business results such as hours saved or faster closes, and a champion who brings only usage to a renewal meeting is easy to overrule. Marketing can help turn activity into a short, credible account of what changed.

The value record is a shared document that customer success keeps and marketing designs: the objective the customer stated when buying, the milestones reached, the result observed and the next goal. Marketing’s part is the template, the plain wording and the one-page summary the champion can forward without editing. If the customer agrees, the same record later becomes a case study through the Case Study Kit.

Running case: Example Co. Fictional

Example Co.’s four value-not-visible losses averaged $8,000 of ARR, more than the $5,500 average across its 20 cancellations. In each, the champion was satisfied, and the chief financial officer cancelled at renewal because nobody had shown what the tool had saved.

Do this

For every account renewing in the next 90 days, write the objective the customer bought for and the result so far in two sentences. If you cannot, that account is at risk whatever its usage says.

When is expansion a next step and not a pitch?

When the customer’s own behaviour shows they need more. Expansion offered on a signal reads as help; expansion offered on a sales target reads as a pitch, and it lands hardest on the accounts that have not yet reached value. These are the signals I look for before any expansion message goes out.

• Sustained adoption. The account has reached activation and used the core workflow steadily for at least a quarter.
• A second team asking. People outside the buying team request access or ask how the product would work for them.
• Usage near a limit. Seats, volume or another plan allowance is close to its ceiling for more than a month.

Running case: Example Co. Fictional

Example Co.’s $200,000 of expansion came from 24 accounts: 16 added a second team at $7,500 each ($120,000) and 8 moved up a volume tier at $10,000 each ($80,000). Every one followed a signal: a second department asked for access, or approvals ran above 80% of the tier allowance for two months. Marketing’s job was to make the next step obvious, with a team-rollout guide and a tier comparison sent only on those triggers.

Pricing decides how much room to expand an account has. If your packaging has no second team, seat or tier to move to, the SaaS Pricing and Packaging Playbook is the place to start, not an expansion campaign.

Do this

Write one expansion trigger per signal, with the message, the owner and the threshold that starts it, and send nothing to accounts that have not reached activation.

What should the first 90 days look like?

Start small: one diagnosis, one friction point, one pilot, measured against a baseline. A retention programme that tries to fix every driver at once rarely shows which change worked.

Running case: Example Co. Fictional

What the pilot is worth to Example Co.: if fit screening and activation work halved the losses from its two marketing-led drivers, $66,000 would become $33,000. Churn and contraction would fall from $150,000 to $117,000, GRR would rise from 85% to 88.3% and NRR from 105% to 108.3%, with expansion held at $200,000. Halving is my assumption for illustration, not a forecast.

Watch the cohort, not the average. The Cohort Retention Curve Builder shows whether customers who signed after the change keep more revenue at the same age than those who signed before. For what better retention does to LTV and payback, see My LTV:CAC Is Below 3:1. Which Number Should I Fix First?

What changes by stage?

• Pre-seed — With a handful of customers, talk to every one of them and write down who is succeeding and why. The success-based ICP starts here, before any targeting is set in campaigns.
• Seed — Define activation, tag every lost customer with one driver and run one 90-day pilot at a time. A founder or the first marketer can own the value record template and the onboarding sequence.
• Series A–B — Give retention marketing an owner who shares the GRR target with customer success. Build role-based onboarding, value records for every renewal and signal-based expansion messages, and report GRR beside NRR to the board.

Method and limits

I opened the SaaS Capital and Bain documents named here on October 29, 2026. The churn-driver table, the four ICP traits, the expansion signals and the 90-day plan are my own working methods, not research findings. Example Co. and all its numbers are fictional, and the arithmetic is shown where each number appears.