How to design a customer marketing motion that drives expansion in 2027

The scenario that forces the redesign

Picture a $40M ARR B2B software vendor heading into planning season. New-logo pipeline is flat, CAC payback has stretched well past the two-year mark that boards used to tolerate, and net revenue retention has drifted from the high 100s down toward breakeven. The CFO does the arithmetic out loud in the QBR: every point of NRR recovered is worth more than a quarter of net-new bookings, and it costs a fraction of the sales headcount. The board asks for an expansion plan by the next meeting.

The customer marketing function in that company, as it exists today, cannot answer. It runs a monthly customer newsletter, coordinates the user conference, chases logos for the website, and fields ad-hoc reference requests from AEs who ping on Slack the day before a late-stage call. It has no number. It has no pipeline. It has no seat in the forecast meeting. When the CRO asks "how much expansion did marketing influence last quarter," the honest answer is nobody tracked it.

That gap — between what the P&L now needs from the install base and what the function is actually built to do — is the reason this redesign exists. It is not a tooling problem first. It is an ownership problem. Expansion revenue lives at the intersection of three teams: customer success owns the relationship and the renewal, sales owns the commercial conversation, and marketing owns the ability to run one-to-many at a scale no CSM book can match. Whoever designs the motion has to design the seams between those three, not just the campaigns.

The same pressure shows up outside pure SaaS, which is worth noting because the pattern generalizes. Managed service providers face it when a client buys one service line and never the second. Payments and fintech platforms face it as merchants stay on the base processing product and never adopt lending, payroll, or spend management. Industrial and equipment vendors face it as service contracts, parts programs, and telemetry subscriptions sit unattached to installed hardware. In every case the economics rhyme: the acquisition cost is already sunk, the relationship already exists, and the constraint is that nobody has built a repeatable mechanism to surface who is ready for more and route them to a human at the right moment.

The scenario also exposes the trap most teams fall into first. Faced with the board ask, the instinct is to buy something — a community platform, an advocacy hub, a fancier email tool — and declare the motion built. Tooling without a routing contract produces campaign activity and zero attributable revenue, which is a worse outcome than doing nothing, because it burns the twelve months of credibility you had to prove the function can carry a number.

How the mechanism actually works

The motion is a signal pipeline with a human handoff in the middle. Strip away vendor names and it has five layers, and each layer answers exactly one question.

The data layer answers "what is true about this account?" A warehouse — Snowflake, BigQuery, Databricks, or a well-governed Postgres at smaller scale — holds product events, billing and invoicing history, CRM opportunity data, and support ticket volume in one place keyed to a stable account ID. This is the unglamorous part that determines whether everything downstream works. If product telemetry uses a workspace ID that nobody has mapped to the Salesforce account ID, your propensity model is fiction. Budget real engineering time for identity resolution before you budget for campaign tools.

The signal layer answers "what is this account doing?" Product analytics — Pendo, Gainsight PX, Mixpanel, Amplitude, or instrumentation you own — turns raw events into usable features: adoption breadth (how many modules touched), adoption depth (how heavily the core module is used), seat utilization against contracted seats, trend direction over the trailing 90 days, and admin-level engagement. Adoption breadth is the single most predictive input in most portfolios, and it is also the easiest to compute, which makes it the right place to start.

The scoring layer answers "who is ready?" You do not need machine learning to start. A weighted rule set beats an unexplainable model in year one because CSMs will actually trust it and argue with it, and their arguments are how the model improves. A workable v1: adoption breadth above a threshold, seat utilization above 80% of contract, a positive 90-day usage trend, an executive relationship logged in the last 60 days, and no open escalation. Score it, threshold it, and call anything above the line an expansion-qualified account.

The orchestration layer answers "what do we send, and to whom?" Marketing automation — HubSpot, Marketo, Customer.io, Braze — runs the one-to-many touches: in-app messaging, lifecycle email, webinar and office-hours invitations, tailored content by industry and product combination. Reverse-ETL tooling (Census, Hightouch, or a scheduled job you maintain) is what moves warehouse-computed scores into the activation tools and the CRM so the same number appears everywhere.

The routing layer answers "who talks to the human?" The customer success platform — Gainsight CS, Catalyst, Vitally, ChurnZero, or a disciplined set of CRM tasks — receives the qualified account as a call-to-action with a stated deadline. This is the layer teams skip and the reason motions die.

The handoff contract deserves its own specification, written down and signed by the VPs before a single play launches. Four clauses, minimum. First, the qualification definition: exactly what score and what conditions make an account qualified, in plain language a CSM can repeat. Second, the disposition SLA: the CSM has a fixed window — five business days is a reasonable default — to accept, defer with a date, or disqualify with a reason code. Reason codes are not bureaucracy; they are your model's training data. Third, the support commitment: once accepted, marketing delivers the supporting assets — a peer reference, an industry-matched case study, a business case template the champion can hand to their own finance team — inside a stated turnaround, typically 72 hours. Fourth, the credit split: expansion revenue is split-credited across CSM, AE, and marketing in the revenue system, so no one is choosing between helping the motion and hitting their own number.

Two things about that loop matter more than the boxes. The disqualification path feeds back into scoring — a motion that never learns why CSMs reject accounts will keep sending the same wrong ones. And the nurture path is not a dead end; accounts that are healthy but not commercially ready are exactly the population your advocacy program should be recruiting from, which means the "no" branch still produces revenue value, just for the new-logo side of the house.

Real numbers, ranges, and the three plays worth building

Three plays produce the overwhelming majority of attributable expansion in most portfolios. Build them in this order, because each one supplies inputs the next one needs.

Feature-launch adoption. Every release is an expansion event, and most companies waste it on a changelog post. The mechanic: instrument the new capability before launch so you can see first-touch and repeat use; define an adoption cohort as accounts whose admins touched the feature in the first 30 days; run a fixed sequence against that cohort — in-app tooltip on first login after release, email at day three with a use-case walkthrough, a CSM nudge at day seven for accounts above a value threshold, and a live office-hours session at day fourteen. Accounts that cross a usage threshold on a feature that sits above their current tier get a commercial prompt rather than another educational email. This play works because it is triggered by behavior the customer already chose, not by a calendar.

Advocacy and reference supply. A named peer reference from a comparable company moves a meaningful share of late-stage enterprise deals, and reference scarcity is a chronic, quantifiable drag on sales cycles that almost nobody measures. Build a rolling pool sized to demand: track how many reference requests sales makes per month, multiply by the fatigue limit you set per customer (three to four calls a year is a common ceiling before goodwill erodes), and size the pool accordingly. Gate entry on objective criteria — adoption breadth, no open escalation, a satisfied executive sponsor, no commercial event in the last quarter — rather than on who the CSM likes. Capture structured outcome data, not adjectives: "cut onboarding from three weeks to nine days" is portable into a case study, a review site, a sales deck, and an analyst briefing. "Great partner" is portable nowhere. The operational metric is reference velocity — net-new referenceable customers added per week — because a pool that doesn't refresh becomes a pool of the same six exhausted logos.

Second-business-unit cross-sell. In multi-product portfolios this is the highest-value play and the hardest, because the buyer for product two frequently sits in a different function, a different region, and sometimes a different budget cycle from the buyer for product one. The mechanic is closer to ABM than to lifecycle marketing: identify accounts using a single product, map the second buying center with contact data and org charts, run a multi-touch sequence that leads with the second buyer's problem rather than your first product's success, and land it with a co-branded executive briefing featuring a peer from the same industry who already runs both. Multi-product accounts retain dramatically better than single-product accounts across essentially every portfolio that publishes the split — the mechanism is switching cost plus workflow entanglement, and it compounds.

On benchmarks, the honest framing is that credible public numbers move and vary hugely by segment, so treat published medians as directional and your own trailing four quarters as truth. The stable, well-documented patterns worth planning against: enterprise-segment NRR runs materially higher than mid-market, which runs higher than SMB, and the spread between segments is usually wider than the spread between good and bad companies within a segment. Expansion typically accounts for a large minority of total new ARR at mature multi-product SaaS companies — often somewhere between a third and a half — and that share rises with portfolio breadth and average contract value. Industry-wide NRR has compressed since the 2021–2022 peak as seat-based expansion slowed and procurement tightened, which is precisely why the motion is being funded now.

Set your own targets from your baseline rather than from a blog post. Pull the last eight quarters of gross retention, expansion ARR, and downgrade ARR by segment. A well-executed motion moving from a standing start reasonably targets a three-to-six point NRR improvement in year one — most of it from downgrade prevention and seat true-ups rather than heroic cross-sell, which lags by two to three quarters because it depends on a second buying center you have not met yet.

Budget honestly. At mid-market scale the tool stack — marketing automation at enterprise tier, product analytics, a CS platform, an advocacy tool, and reverse-ETL — lands in the low-to-mid six figures annually before headcount. Vendor list pricing for this category is negotiated heavily and published prices are close to meaningless at contract time; assume a meaningful discount off list on multi-year commitments and budget for the implementation services, which routinely run a third of first-year license cost. Staffing at that scale is typically one leader plus two ICs plus a fractional RevOps analyst. The expensive line is not software. It is the data engineering to make account identity reliable across five systems, and teams underestimate it every single time.

Trade-offs, alternatives, and what to build instead

Every design choice here has a live alternative, and the right answer depends on your portfolio shape, not on best practice.

Marketing-owned versus CS-owned expansion. If your average contract value is high and your CSM book is small — say, twenty accounts per CSM — a human-led motion outperforms anything marketing can automate, and customer marketing's job shrinks to supplying content, references, and business-case artifacts. If books run to a hundred-plus accounts, or you have a long tail of self-serve customers no human touches, marketing-led one-to-many is the only mechanism that reaches them. Most companies have both populations and need both motions, segmented explicitly. The failure is running one motion across both and wondering why enterprise CSMs ignore your emails while the tail goes unserved.

Rules versus models for propensity. A transparent weighted rule set ships in weeks, gets argued with productively, and is defensible to a CSM who disagrees. A trained model finds interactions humans miss but needs a year of labeled outcomes and dies on adoption if nobody can explain a score. Start with rules, log every disposition with a reason code, and revisit modeling once you have four quarters of labels. Skipping the rules phase means skipping the labels phase, which means the model you eventually want has nothing to learn from.

Build versus buy on the CS platform. A dedicated CS platform is the right call once you have more than a handful of CSMs, multiple playbooks, and health scores that drive real routing. Below that, CRM tasks plus a warehouse view plus disciplined process genuinely works and costs nothing, and the discipline you build manually transfers cleanly when you do buy. Buying a platform to create process you don't have yet produces expensive shelfware.

Community as a growth lever. Community platforms are attractive and slow. They rarely show attributable retention impact inside the first year, they need dedicated staffing to avoid becoming a ghost town, and they compete for the same budget as the three plays above. The exception worth naming: if your product has a genuine practitioner-identity dimension — where users build careers around your certification and want to be seen in that context — community can outperform everything else. Know which of those you are before signing.

Attribution model. Split credit invites gaming, and single-owner credit invites hoarding. Most teams land on split credit with a materiality floor: below a certain deal size, don't bother splitting, because the operational overhead exceeds the behavioral benefit.

Pitfalls that kill the motion, and the early warning for each

Launching before the routing contract is signed. The most common failure and the most preventable. If a CSM learns about a qualified account by being copied on a marketing email, the motion is dead inside two months — not because the CSM is territorial, but because you've told them their account relationship is something you'll operate around rather than through. Early warning: disposition rate below 60% in the first month. Fix: co-design the SLA with CS leadership, co-present it at kickoff, and route every qualified account through the CS tool before any customer-facing touch fires.

Confusing sentiment with intent. High satisfaction scores and high NPS do not predict purchase. Plenty of delighted customers are fully served by what they already own, and plenty of grumpy ones are expanding because the product is load-bearing. Sentiment is a churn-risk input, not an expansion-propensity input. Early warning: your top-scoring accounts by propensity are the same list as your top NPS promoters — that means your model is reading survey data, not behavior. Fix: weight usage breadth and depth, and treat sentiment as a suppression filter rather than a driver.

No shared dashboard. When the CRO reads one number in the forecast tool, CS reads another in the health platform, and marketing reads a third in the automation tool, the weekly meeting becomes a reconciliation exercise and the motion loses its sponsor. Early warning: the first ten minutes of the expansion standup are spent arguing about whose number is right. Fix: one dashboard, one definition of expansion ARR, one owner of the definition — RevOps — and a written data dictionary.

Compensation that contradicts the design. If the CSM is paid purely on gross retention, expansion conversations are unpaid work that risks the renewal, and a rational CSM will deprioritize them. If the customer marketing leader has no variable component tied to revenue, the function drifts back toward events and swag within three quarters because that's what gets visibly appreciated. Early warning: qualified accounts being deferred with vague reasons. Fix: put a meaningful slice of variable comp on a blended retention-plus-expansion goal for both roles, and make sure finance can actually calculate it before you publish the plan.

Signal quality nobody audits. Product instrumentation rots. A release renames an event, a workspace migration breaks the account mapping, an integration silently stops syncing, and your scores quietly become noise while the dashboard keeps rendering. Early warning: qualification volume changes more than 30% week over week without a corresponding business event. Fix: a monthly signal audit with row-count and distribution checks on every input, treated with the same seriousness as a financial close.

Over-firing at the customer. Lifecycle marketing, CS outreach, product in-app messages, renewal notices, and the newsletter all hit the same admin, and no single system sees the total. The champion's inbox is the shared, finite resource this whole design spends. Early warning: unsubscribe rates climbing among your highest-value accounts — the exact population you can least afford to lose contact with. Fix: a global frequency cap enforced across marketing automation, the CS platform, and in-app messaging, with commercial touches taking priority over educational ones during an active expansion conversation.

Declaring victory on activity. Campaigns sent, webinars run, and advocates recruited are inputs. If the quarterly review reports those instead of expansion ARR, qualified-account conversion rate, and time-to-second-product, the function has quietly reverted to a comms team with better tooling. Early warning: your own slide has no dollar figure on it.

Related questions

How long before the motion shows measurable revenue?

Expect first qualified accounts within 90 days, first closed expansion attributable to the motion around month four to five, and a readable NRR trend at month nine to twelve. Cross-sell to a second buying center lags longest because it requires relationships you haven't built yet.

Who should own the propensity score definition?

RevOps owns the definition and the data dictionary; customer marketing proposes changes based on play performance; CS has veto over thresholds that would flood their queue. Shared authorship without a single owner produces a score that drifts every quarter and that nobody trusts.

Does this work for companies without product telemetry?

Yes, with weaker signals. Substitute support ticket themes, login frequency, invoice and overage patterns, seat utilization against contract, and structured QBR notes. The motion architecture is identical — the scoring layer is just coarser, so lean harder on CSM judgment in the disposition step.

How is this different from a standard renewal motion?

Renewal defends existing revenue on a contract calendar. Expansion adds revenue on a behavioral trigger. They share a customer and often a CSM, but conflating them means expansion conversations only happen in the 90-day renewal window, which is the worst possible moment to introduce new spend.

What is the smallest viable version of this?

One warehouse view of adoption breadth by account, one weekly list of the top twenty accounts crossing a threshold, one CSM disposition meeting, and one shared spreadsheet tracking outcomes. Prove the conversion rate before buying anything.

FAQ

What exactly is customer marketing responsible for in an expansion motion?

Customer marketing owns the one-to-many layer: identifying which install-base accounts are ready for more, running the lifecycle programs that build adoption and educate on adjacent products, supplying references and business-case artifacts to accelerate deals, and reporting the resulting pipeline. It does not own the commercial conversation or the relationship — those stay with sales and customer success respectively.

Do we need a data warehouse before we start?

Not to start, but you will need one to scale past the first play. A single reliable join between product usage and CRM account records is the actual requirement; that can live in a warehouse, a well-maintained CRM sync, or even a scheduled export in the earliest phase. What you cannot skip is a stable account identifier that means the same thing in every system.

How do we avoid annoying customers with expansion outreach?

Trigger on behavior rather than calendar, enforce a global frequency cap across every system that can message an admin, lead every touch with a use case rather than a product pitch, and give customer success the authority to suppress any account with an open issue. The strongest signal you're getting this right is that champions forward your emails internally instead of unsubscribing.

Should the customer marketing leader carry a revenue number?

Yes, if you want the function to survive as a revenue function. A blended goal on retention plus expansion ARR, with a meaningful variable component, is what keeps the design intact when the next user conference eats the calendar. Without it, the role reliably drifts back to events and content within a year.

What is the single highest-leverage thing to build first?

The routing contract between marketing, customer success, and sales — before any tool purchase. A mediocre play with a signed handoff SLA outperforms an excellent play that has nowhere to land. Write down the qualification definition, the disposition window, the support turnaround, and the credit split, and get both VPs to sign it.

How do we handle accounts that are healthy but genuinely have nothing left to buy?

Route them to advocacy. A fully-penetrated happy account is your most valuable reference supply, and treating that as a real outcome rather than a dead end keeps the nurture branch of the motion economically productive. Track their contribution as reference velocity feeding the new-logo pipeline, and credit the team for it.

Sources

• ChartMogul SaaS Retention Report — net revenue retention distributions and trend data across B2B SaaS cohorts
• SaaS Capital — Private SaaS Company Growth Rate Benchmarks — annual survey data on growth, retention, and efficiency at private B2B software companies
• Gartner — B2B Buying and Sales Research — buying-group behavior, buyer-enablement content, and reference influence in complex deals
• Forrester Research Blogs — customer advocacy technology category analysis and post-sale marketing coverage
• Bain & Company — Net Promoter System — origin and appropriate use of NPS, including its limits as a revenue predictor
• Harvard Business Review — Customer Retention and Growth — research on retention economics and the cost asymmetry between acquisition and expansion
• OpenView Partners — Product Benchmarks archive — historical product-led growth and expansion benchmark reporting
• Pavilion — GTM benchmark research — operator-sourced go-to-market benchmarks across SaaS segments
• G2 — Customer Advocacy Software category — vendor landscape, feature comparison, and verified user reviews
• Gainsight — Customer Success resources — health scoring, playbook design, and CS-to-sales handoff practice

Related on PULSE

• [CSM-Led Expansion Motion Architecture in 2027](/knowledge/ra0410)
• [ABM Org Structure: Sales + Marketing Pod Design in 2027](/knowledge/ra0244)
• [How to build a Revenue Council across Sales/Marketing/CS/Finance in 2027](/knowledge/ra0318)
• [Product Marketing Org Structure for Multi-Product SaaS in 2027](/knowledge/ra0245)
• [How to run a quarterly Sales Kickoff that drives behavior change in 2027](/knowledge/ra0307)