Why your installed base is the next profit frontier

1. The asymmetry hiding in plain sight

Most B2B growth plans read like they were written backwards.

Open the strategy deck of almost any revenue, marketing or product leader, and the center of gravity is net-new logos: pipeline targets, demand-gen spend, win-rate improvement, sales-cycle compression. The installed base — the customers and assets you have already won — usually appears later, in a slide called “Retention” or “Customer Success,” with a churn number attached and not much else.

The economy does not justify that ordering.

Retaining a customer costs less than a third of acquiring one, and on average existing customers generate around 10% more revenue than new ones (Bates et al., 2025). In industrial settings the multiple is sharper still: aftermarket operating margins run at roughly 2.5× the margins on new equipment, and 40–50% of total profit in many manufacturers now comes from services delivered to the installed base rather than from new unit sales (Millar, Bendig and Hussain, 2020). The pattern is structural, not anecdotal: services are typically sold to a customer base the supplier already knows, can monitor, and can reach without a fresh acquisition cost (Egleston et al., 2022).

The pattern repeats in software. McKinsey’s analysis of 98 B2B SaaS companies found that top-quartile firms on Net Revenue Retention (NRR) achieve NRR of around 113%. This means they grow 13% a year from existing customers alone, before a single new logo closes while bottom-quartile firms sit at 98% and slowly shrink (Pulido, Erdogan and Berliner, 2025). The gap between the two cohorts compounds dramatically over three to five years. It is also durable: top-quartile NRR firms sustain higher valuations than peers through both bull and bear markets, and outperform on payback period and growth efficiency at the same time.

The foundational economics have been known for thirty-five years. Reichheld and Sasser (1990) showed that reducing the defection rate by just 5 percentage points generated 30–85% more profit depending on industry: (1) 85% in one bank’s branch system, (2) 50% in an insurance brokerage, (3) 30% in an auto-service chain. Reichheld (1996) extended this into a full model of loyalty economics that consultancies still build on today.

None of this is new. What is new is that customer acquisition has become materially harder and more expensive at exactly the moment when installed-base economics have become easier to instrument. Acquisition is now slower (sales cycles have lengthened across every B2B segment), more crowded (every category has a longer vendor list), and digitally noisier (buyers actively avoid suppliers who send irrelevant outreach). Meanwhile, the data infrastructure required to systematically activate the installed base — usage telemetry, CRM-CS integration, AI-driven prioritisation — has moved from aspirational to standard.

The marginal dollar of growth investment in most B2B organisations is misallocated. That is the asymmetry. The rest of this article is about what to do with it.

2. Why your installed base is really a data asset

The instinct, when a leader is told to “grow inside the base,” is to reach for the customer list and start segmenting accounts. That instinct is wrong, or at least incomplete. The list is the smaller half of the asset.

Ulaga and Reinartz (2011), in the canonical academic study of how manufacturers combine goods and services profitably, identified four critical resources for successful expansion. The first, listed before the field service organisation, before the existing sales force, before the product development of assets, was “product usage and process data derived from the firm’s installed base of physical goods.” The point generalises beyond manufacturing. The installed base is primarily a data asset. it is the only place in a B2B business where you can see what customers do, not just what they bought.

That distinction matters because it changes what the activation work looks like.

If the installed base is a list, the work is sales coverage: assign accounts, set quotas, run Quarterly Business Reviews (QBR). If the installed base is a data asset, the work is signal architecture: identify which behavioral and contextual data predicts which expansion motion, then build the prioritisation engine that connects signal to action.

McKinsey’s industrial-aftermarket work operationalises this through the concept of “gap to entitlement”.  The proportion of a customer’s potential aftermarket spend that the supplier is not currently capturing. One industrial player with limited installed-base visibility scraped external data, enriched its transactional database, and used analytics to estimate entitlement share by product line. The exercise surfaced a 15–25% potential revenue uplift and produced distributor-level scorecards that drove performance conversations (Egleston et al., 2022). A global medtech business used propensity-to-buy models to direct an inside sales team toward customers most likely to convert maintenance contracts (built from sub-sector, service history and buying behavior). Their overall sales rose by 15%.

The same logic is now standard in SaaS, just with different vocabulary. Pulido et al. (2025) found that “product telemetry” was one of nine practices that materially separated top-quartile NRR firms from the rest. The leaders had near-real-time dashboards that let them drill into micro-segments and see, for instance, how a price increase or the end of a promotion affected actual customer behavior. The laggards either did not have the telemetry or did not have the practice of acting on it.

The implication for product and revenue leaders is uncomfortable but clarifying: the installed-base opportunity is bounded less by sales capacity than by signal quality. A team that knows which 200 accounts are most likely to expand next quarter, why, and into what, can run a smaller, more focused motion than a team that has the customer list and a quota.

3. The four signals that predict expansion

If signal quality is the bottleneck, the practical question is which signals matter. The literature does not give a single tidy answer; the implementations described in McKinsey’s aftermarket work (Egleston et al., 2022) and the NRR-practice research (Pulido et al., 2025) suggest four signal families, each tied to a different expansion motion.

Usage intensity is the most familiar. In SaaS, it is feature adoption, depth of use, frequency of login, and share of available seats activated. In industrial settings it is machine hours, throughput, cycle counts, and consumable replenishment cadence. Usage data predicts both expansion opportunity (a customer hitting 80% of tier limits is a near-term upgrade) and churn risk (a customer whose usage has flattened for two quarters is at risk regardless of NPS). The two share an underlying mechanism: customers who use the product more are both more likely to expand and less likely to leave. Pulido et al. (2025) found that customers who expand are also 60–70% less likely to churn than customers who remain flat.

Context shifts are less instrumented but often more predictive. Renewals are the obvious context event. Less obvious: a customer’s leadership change (new CFO, new VP of Operations), a funding round, an acquisition, a restructuring, a new product launch by the customer themselves. These shifts re-open buying windows that had been closed; they also create the political cover to expand a vendor relationship that an incumbent buyer was previously content to leave alone. Most CRMs capture none of this systematically.

Buying signals sit between the two. Service ticket patterns, support escalation themes, search behavior on the supplier’s own properties, RFP activity, attendance at supplier events, increased engagement from a wider buying committee — each is a small piece of evidence on its own and a strong leading indicator in combination. Egleston et al. (2022) describe a multinational machine-tool builder that combined CRM data, shipment data and machine-learning models into an automated outreach stack. This generated 20% more leads and freed 20% of the inside-sales team’s bandwidth for higher-value work.

White space is the structural signal. For any account, it is the difference between what the customer could buy across the supplier’s full portfolio and what they actually do. A B2B packaging company described by Bates et al. (2025) achieved a tenfold increase in cross-sell revenue in eight weeks by sizing share of wallet for key customers, mapping white space analytically, and running hyper-personalised marketing for the priority accounts. The signal was not new, but the discipline of using it was.

The original contribution available to most organisations is not in choosing among these four, it is in unifying them. The four signal families typically live in four different systems, owned by four different functions, with four different update cadences. The team that integrates them first wins.

4. The motion: who does what, when

A unified signal architecture does nothing if the commercial motion that consumes it is still designed around net-new acquisition. The activation system has to be redesigned downstream too.

The dominant industry framework for post-sale coverage has been the “high-touch, medium-touch, tech-touch” pyramid. It does the following: tier customers by size and allocate human attention accordingly. Pulido et al. (2025) describe this model as “still necessary, but no longer sufficient.” Their survey found that only 3% of B2B SaaS respondents had best-in-class coverage models. The models are:

• dynamic,

• segmented by customer need rather than just by size,

• and capable of moving accounts between coverage modes as their signals changed.

One IT service-management vendor running this dynamic model achieved an NRR of 125%, far above the 113% top-quartile benchmark.

The replacement model has three properties.

1. Segmentation by need, not by ACV.  A small-but-strategic account in a fast-growing customer industry may warrant more touch than a large account in maintenance mode.

2. Signal-triggered movement. Accounts shift between high-touch and digital-led coverage based on usage, context and buying-signal data rather than annual planning.

3. Clear ownership of each NRR component. Retention, pricing and expansion are separated, owned by different leaders with different incentives, and re-integrated at the CEO/CFO level.

The third property is the one most organisations get wrong. Pulido et al. (2025) note that “it is rare, and often not possible, for one leader to be responsible for the entirety of the NRR metric,” because retention typically sits under customer success, pricing under product or finance, and expansion under sales. TSIA’s 2025 benchmark research describes the consequence in field terms: low-complexity expansions  (small add-ons, service attaches, seat increases ) slip through the gaps between sales and customer success teams because neither function is unambiguously accountable for them, and neither finds them economically interesting at the individual deal level (Technology & Services Industry Association, 2025). At the system level the lost revenue is substantial.

The TSIA data also pushes back on a common organisational reflex. Several years of benchmark data show that giving net-new sales executives responsibility for renewals and low-complexity expansion reduces growth and increases retention costs, while making no measurable difference to renewal rates. The land sales motion and the expand-and-renew motion require different capabilities, different cadences and different incentive structures. Mixing them costs both.

The more productive question is not who closes expansion deals but who owns the framework that makes expansion systematic. The evidence is not yet decisive, but the most defensible answer is that product managers and product-line leaders with P&L responsibility for the product (the people who already own the value proposition, the pricing architecture, and the usage data) are the natural owners. Sales executes new logo. Customer success executes adoption and renewal. The PM/PLL owns the activation framework that connects the two and ties it to product economics. CMOs and CSO/CROs share accountability for the orchestration.

This is an editorial position, not a published finding. Reasonable practitioners will disagree on the org chart. What is harder to disagree with is that fragmented ownership predicts under-performance, and that the fragmentation is the rule rather than the exception.

5. The pricing and packaging layer no one wants to fix

Of the nine practices Pulido et al. (2025) identified as material differentiators of NRR outperformance, the largest single lever was not customer success, not telemetry, not coverage model. It was pricing and packaging discipline. Firms with best-in-class practices in this area achieved NRR roughly 16 percentage points higher than peers with only basic practices in place.

This is the part of installed-base activation that most organisations would prefer to skip.

The discipline has six elements.

1. Product pricing and packaging designed with multiple predefined upsell paths, so growth in customer value translates automatically into manufacturer revenue through capacity tiers, modules and add-ons.

2. Cross-sell bundles that incentivise platform consolidation.

3. In-product journeys that drive trial and adoption of new functionality.

4. Standard terms and conditions with explicit penalties for partial cancellation, reinstatement and license abuse that are enforced and not aspirational.

5. Discount policies that restrict ad-hoc concessions and automatically reprice when a customer drops a product from a multiproduct bundle.

6. Automated quoting of the great majority of expansion and renewal opportunities through CPQ tools integrated with the CRM, supported by dynamic deal-scoring models that give granular discounting guidance.

Each element is operationally unglamorous. Together they are the difference between an installed base that compounds and one that leaks.

The academic literature reinforces this in a more uncomfortable form. Eggert et al. (2011), in a longitudinal study of 414 German mechanical engineering firms, distinguished Services Supporting the Product (SSP : maintenance, parts, repair) from Services Supporting the Client (SSC: process consulting, performance management, outcomes). Both increased revenue level and growth. But they had differential effects on profit, and the higher-value SSC motion required deeper organisational change. Fang, Palmatier and Steenkamp (2008) found, in their landmark study of service-transition strategies, that firm value rose with services as a share of total sales, but only once the share crossed a threshold of roughly 20–30%. Below that threshold, the transition was a cost, and above it, the transition compounded.

The implication is that activation is not a marketing exercise. It is a pricing, packaging and contract-architecture exercise that happens to require marketing, sales and customer success to execute. Most of the organisational design effort should be at the seam between product management and finance, not at the seam between sales and marketing.

6. The honest counterpoint: where this goes wrong

Every consulting deck about installed-base activation features the same hockey-stick chart. The academic literature features a different shape.

Gebauer, Fleisch and Friedli (2005) named it the service paradox. Manufacturers invested in extending their service business to the installed base, incurred the higher costs of service provision, and then failed to realise the returns. Not occasionally but repeatedly, across enough firms to constitute a pattern. The paradox is not that installed-base economics are bad. It is that the transition destroys value in firms that change their offering without changing their organisation, their incentives and their culture.

Neely (2008) provided the most uncomfortable empirical data on this. His analysis of 10,028 firms across 25 countries found that servitized firms, those that had expanded into the installed base with services and outcomes, were systematically larger in revenue terms but generated lower profits as a percentage of sales than traditional manufacturers. The sample also contained a higher proportion of bankrupt servitized firms than would be expected from base rates. Servitization, in other words, was correlated with revenue growth but not profit growth at the cross-sectional level, and a non-trivial minority of firms destroyed value through the transition.

Eggert et al. (2011) gave the dynamic version. In their longitudinal data, industrial-service strategies increased both the level and the growth of revenues, as expected. But they reduced the level of profits before improving profit growth. The J-curve is real. There is a margin trough before the compounding starts, and the trough is long enough that boards and CFOs who have not been prepared for it will pull the plug at month 12 or month 18, exactly when the model says they should hold their course.

Ambroise, Prim-Allaz and Teyssier (2018) added the configurational nuance. Financial performance from servitization does not depend on the strategy choice in isolation. It depends on the fit between the strategy and the firm’s customer-oriented organisational design (its service culture, customer interface, and service delivery system). In their 184 firm sample, the firms that succeeded were the ones whose organisational design matched their strategic ambition. The firms that failed had picked a strategy and left the organisation unchanged.

The lesson for the article’s argument is not that installed-base activation is wrong. The lesson is that the case for it is not “this is easy money you have been leaving on the table.” The case is “this is structurally higher-quality growth than acquisition, and the transition is harder and longer than the consulting case implies, and a meaningful minority of firms that attempt it will fail.” Both halves of that sentence are required.

The recent SaaS data reinforces the point in a different way. Achieving NRR above 100% has become measurably harder over 2022–2024 across most ARR bands as growth has slowed, churn has crept up, and consumption-based pricing has exposed suppliers to customer-demand fluctuations they do not control (Pulido et al., 2025; Technology & Services Industry Association, 2025). The compounding still works and it just requires more deliberate engineering than it did in 2018.

7. Measurement and the road from here

If activation is going to be defensible inside the company, the metrics have to change.

Two shifts matter. The first is the move from self-reported customer-satisfaction scores to accounting-based measures of customer-driven growth. Reichheld, Darnell and Burns (2021) introduced Earned Growth Rate as the auditable counterpart to Net Promoter Score. It is the share of revenue growth that comes from returning customers and their referrals, calculated from internal accounting data rather than survey responses . The shift matters because NPS, despite being used by two-thirds of the Fortune 1000, has been “gamed and misused” in ways that reduced its credibility, in the words of its own inventor. Earned Growth Rate is harder to manipulate because it is derived from audited revenues.

The second shift is the disaggregation of NRR. A single NRR number combines retention, pricing, and expansion. These three motions have different owners, different drivers and different time horizons. Companies that manage NRR well track the three components separately and tie each one to a clear leader with clear incentives. Pulido et al. (2025) describe this as building an “NRR MVP early-warning backbone”. It comprises of predictive analytics focused on the highest-value use cases (adoption, retention, cross-sell) rather than a sprawling business-rules engine that nobody trusts.

Both shifts depend on a third, more cultural change: instrumenting the J-curve. If the activation transition has a margin trough before the compounding starts, the CFO needs to see the trough modelled and approved before the program begins. The management team must know when revenue uplift and margin recovery are expected, and what intermediate signals will confirm or disconfirm the model. Without that, every activation program is one missed quarter away from being killed.

The installed base is not a retention problem. It is a growth engine that has been undermanaged, because it has been measured as a cost. The data infrastructure that makes systematic activation possible is now affordable. The commercial motion that converts data into expansion is well-described. The pricing and packaging architecture that compounds the result is unglamorous but tractable. The honest counterpoint, that the transition is harder than the slide deck suggests, and that a minority of firms will fail, is exactly why competitive advantage is available to the firms that get the transition right.

The asymmetry is real. The marginal dollar of growth investment in most B2B organisations is misallocated. The opportunity is to move it before everyone else does.

References 

Ambroise, L., Prim-Allaz, I. and Teyssier, C. (2018) ‘Financial performance of servitized manufacturing firms: A configuration issue between servitization strategies and customer-oriented organizational design’, Industrial Marketing Management, 71, pp. 54–68.

Bates, D., Mirza, Y., Medha, M. and Lamoureux, T. (2025) ‘Growth amid uncertainty: Jump-starting B2B sales performance’, McKinsey & Company, 27 August.

Eggert, A., Hogreve, J., Ulaga, W. and Muenkhoff, E. (2011) ‘Industrial services, product innovations, and firm profitability: A multiple-group latent growth curve analysis’, Industrial Marketing Management, 40(5), pp. 661–670.

Egleston, E., El Khoury, G., Lavandier, H., Muthiah, S. and Zegeye, A. (2022) ‘Five digital and analytics battlegrounds for B2B aftermarket growth’, McKinsey & Company, 17 February.

Fang, E., Palmatier, R. W. and Steenkamp, J.-B. E. M. (2008) ‘Effect of service transition strategies on firm value’, Journal of Marketing, 72(5), pp. 1–14.

Gebauer, H., Fleisch, E. and Friedli, T. (2005) ‘Overcoming the service paradox in manufacturing companies’, European Management Journal, 23(1), pp. 14–26.

Millar, K., Bendig, O. and Hussain, A. (2020) ‘Aftermarket services: Transforming manufacturing in the wake of the COVID-19 pandemic’, Deloitte Insights, 14 May.

Neely, A. (2008) ‘Exploring the financial consequences of the servitization of manufacturing’, Operations Management Research, 1(2), pp. 103–118.

Pulido, A., Erdogan, B. and Berliner, J. (2025) ‘The net revenue retention advantage: Driving success in B2B tech’, McKinsey & Company, 19 November.

Reichheld, F. F. (1996) The Loyalty Effect: The Hidden Force Behind Growth, Profits, and Lasting Value. Boston, MA: Harvard Business School Press.

Reichheld, F., Darnell, D. and Burns, M. (2021) ‘Net Promoter 3.0’, Harvard Business Review, 99(6), November–December, pp. 78–87.

Reichheld, F. F. and Sasser, W. E. Jr. (1990) ‘Zero defections: Quality comes to services’, Harvard Business Review, 68(5), September–October, pp. 105–111.

Technology & Services Industry Association (2025) The State of Customer Growth and Renewal 2025.

Ulaga, W. and Reinartz, W. J. (2011) ‘Hybrid offerings: How manufacturing firms combine goods and services successfully’, Journal of Marketing, 75(6), pp. 5–23.