Turn the installed base into an aftermarket pipeline

Most OEMs can find signs of demand. A service engineer sees a worn component. A warranty approaches expiry. A controller reaches end of support. A connected machine reports abnormal vibration. A customer buys far fewer wear parts than comparable sites. The signs sit across service reports, asset records, engineering notices, parts orders, contracts and conversations.

The hard part is turning those signs into credible work. Sales cannot call every old machine an opportunity. Service cannot flood CRM with alarms. Product teams cannot launch an upgrade without checking serial ranges, configuration and supply. The customer needs a useful recommendation tied to uptime, safety, capacity, cost or compliance.

This guide builds that path from active equipment to booked and delivered aftermarket work. It covers triggers, eligibility, evidence, ownership, stage gates, forecasts, suppression rules and the feedback that updates the installed machine after delivery.

Define the pipeline at asset level

The basic record should describe one need on one asset or a controlled group of equivalent assets. It needs an asset ID, customer and site, trigger, affected configuration, proposed action, evidence, value to the customer, owner, channel, timing window and current stage. A fleet campaign can contain thousands of records, but each item still needs a machine-level outcome.

This unit prevents three common errors. The team does not count one account as one opportunity when it owns 200 machines. It does not send an offer for a retrofit that fits only twelve serial numbers. It does not lose the relationship between the quote and the equipment that should change after delivery.

Use an account opportunity above the asset records when the customer wants one negotiation or contract. Retain the asset items underneath it. The commercial roll-up should never erase technical eligibility or execution status.

Start with the active fleet you can serve

Choose the population before searching for opportunities. Confirm which machines remain active, where they operate, who owns and uses them, which company has service rights and whether your organization can still supply the required parts, skill and documentation. Mark unknown status instead of assuming every historical shipment still runs.

Installed-base size drives the available market. Its composition drives the offer. Age, duty, configuration, operating environment, service history and current coverage change what each machine needs and when it needs it.

Research on industrial service sales forecasting found that installed-base characteristics can help OEMs predict active customers and service volume, but the model depends on comprehensive, current installed-base data. Start with a bounded fleet whose status the team can verify. Expand as field work, customer contact and order history improve the record.

Create a trigger library

A trigger states why the team should look at an asset now. Define each trigger with its source, threshold, evidence, eligible offer, review cadence, expiry and suppression rule. This turns a vague idea such as “old machine” into a repeatable test such as “drive family D4, at least nine years in service, original controller installed, support ends within twelve months.”

Use several trigger families because aftermarket demand does not come from telemetry alone. Time, contract, engineering, service, parts, customer and commercial records each reveal different needs.

Treat a signal as a prompt to investigate

A signal says something changed. It does not prove a sale exists. A high-temperature alert may come from a bad sensor, a known production event or a duplicate message. A warranty date may refer to a retired machine. Low parts purchases may reflect light duty, third-party supply or inaccurate asset ownership.

Preserve the source event and its timestamp. Group duplicate signals, apply a review window and link new evidence to the existing case. Microsoft documents this pattern in Dynamics 365 Field Service: rules create an IoT alert, similar alerts can share a parent and a reviewed alert can become a case or work order. The same discipline applies to commercial triggers. One machine condition should create one controlled investigation.

Set a minimum evidence threshold for each trigger. A contract-expiry trigger can qualify from an active asset, valid contract and confirmed expiry. A predictive-failure trigger may need several readings, a diagnostic rule and a technical reviewer. Stronger claims need stronger evidence.

Check technical eligibility before customer outreach

The technical gate answers whether the proposed work applies to the installed configuration. Check model, serial range, revision, as-maintained components, software or firmware, interfaces, duty, environment, regulatory region and prior modifications. Confirm prerequisites, exclusions and the result after installation.

Product engineering should publish an eligibility rule for repeatable offers. The rule can return eligible, ineligible or engineering review. It should cite the bulletin, service instruction, compatibility table or approved configuration that supports the result.

A field photo, controller export or nameplate may resolve missing data. Ask for the smallest piece of evidence that settles the decision. Do not send a salesperson to discover during the customer call that the upgrade cannot fit the machine.

State the customer reason to act

Translate the trigger into an operating consequence the customer recognizes. Connect the recommendation to uptime, safety, quality, energy, throughput, labor, supportability, compliance or total cost. Use the machine history and customer duty instead of a generic campaign claim.

For a controller retrofit, the reason might be the combination of an unsupported component, a twelve-week replacement lead time and a critical production line. For a service agreement, it might be a remote site, rising corrective visits and no local spares. For a wear-parts program, it might be missed planned changes that cause secondary damage.

Show the evidence and its limits. If the OEM can calculate avoided downtime from the customer’s own history, use it. If the team only knows that failure exposure increased, say so and propose an inspection. Credibility creates the next conversation.

Score need, readiness and fit separately

One combined score hides the reason behind the rank. Use separate views for technical need, customer impact, timing, commercial access, offer readiness and data confidence. Apply hard gates for safety, eligibility, entitlement and channel rights before ranking the remaining candidates.

A technically urgent asset can have low contact readiness because the dealer owns the relationship. A high-value upgrade can have low offer readiness because engineering has not released the kit. A likely renewal can have weak data confidence because the asset changed owners. Those cases need different work.

Find the buyer and respect the route to market

Map the asset owner, operator, maintenance lead, reliability lead, procurement contact, budget owner and site approver. The person who reports the condition may not control the purchase. The corporate account may negotiate a framework while the plant funds the work.

Record distributor, dealer and service-partner rights before outreach. Decide who owns the relationship, who prepares the technical offer, who quotes, who delivers and who receives credit. A direct OEM campaign that bypasses an authorized channel can damage more value than it creates.

For fleet customers, group asset needs into a site plan or corporate program when that makes the decision easier. Retain the machine list, eligibility and execution state under the commercial proposal.

Choose the right commercial motion

Different needs require different paths. A safety action may begin with controlled communication and mandatory work. A routine maintenance due date can create a work-order proposal. A large upgrade may require discovery, engineering validation, a budgetary estimate and a planned shutdown. A coverage renewal should enter the customer’s budget cycle months before expiry.

Choose among service recommendation, parts quote, inspection, repair or exchange, service agreement, retrofit, modernization, software update, training and replacement equipment. The cheapest offer may fail to address the customer’s real exposure. The broadest offer may delay an urgent repair.

Define the entry product and the expansion path. A paid inspection can establish scope for an overhaul. A parts order can reveal a stocking need. A repair history can support an upgrade. Each step should deliver value even if the customer stops there.

Build an offer package the field can trust

A repeatable aftermarket offer needs released technical scope, applicability, exclusions, prerequisites, required parts, labor, tools, skill, duration, lead time, documentation, warranty, price logic and expected installed-state change. Give sales and service one approved package with controlled options.

Confirm supply before multiplying the campaign. An obsolete part notice without a viable substitute creates anxiety. An upgrade campaign without kits or trained technicians creates a backlog. Limit the qualified population to what the OEM and its channel can deliver in the promised window.

Salesforce describes a related structure in Manufacturing Cloud product service campaigns: teams segment affected assets, add them to a campaign and track recall, upgrade, repair or other work. The software record matters less than the discipline of connecting campaign scope, asset eligibility and execution.

Use stage gates that prove progress

Stages should describe evidence, not optimism. Define the condition that opens and closes each stage. Assign an owner and maximum age. Do not let a generated signal inflate the sales forecast.

Do not turn every asset signal into a CRM opportunity

Maintain the early population and signals in an asset or campaign workbench. Create a commercial opportunity when the team has technical fit, a plausible customer reason, an accountable route to market and a defined next action. This protects sales attention and forecast quality.

A rejected candidate still has value. Store the reason: inactive asset, wrong configuration, customer self-maintains, third-party coverage, no buyer access, supply unavailable, duplicate, poor economics or customer declined. Set an expiry or review date where the condition can change.

Suppress repeat outreach after a clear answer. A customer who declined a nonurgent upgrade last month should not receive the same recommendation from another region next week. Reopen the item only when new evidence changes the case.

Forecast from the work remaining

Pipeline amount should reflect qualified assets, offer value, customer timing and a stage probability based on actual history. Separate detected potential from sales forecast. Report both, with clear definitions.

Estimate population value from active eligible assets and the appropriate offer. Estimate near-term qualified value from assets that pass the technical and commercial gates. Estimate forecast from customer-confirmed opportunities. Capacity, parts and channel limits can cap the executable amount even when demand exceeds it.

Use different conversion curves by trigger and offer. A service-agreement renewal behaves differently from an unsolicited modernization. A safety campaign behaves differently from a consumables plan. Backtest by cohort, age and region before changing probabilities.

Prioritize customer value and delivery capacity together

Rank urgent safety and uptime exposure first. Then consider customer impact, probability, contribution, relationship importance, supply position and service capacity. A high-margin campaign that consumes a scarce engineering team can displace more valuable work.

Reserve parts, labor and partner capacity as the opportunity becomes firm. Use staged reservations so early pipeline does not block urgent service. Release reservations when the decision date passes or the customer declines.

Coordinate campaigns that hit the same site. A customer should receive one coherent equipment plan, not separate calls for warranty, controls, parts and service from four OEM teams.

Use field service as a source of qualified evidence

Technicians see configuration, operating conditions, competitive equipment, user behavior and unresolved needs that central records miss. Give them a short way to capture the asset, observation, evidence and recommended follow-up. Ask for a photo, reading or service finding when it helps qualification.

Do not turn every technician into a quota-carrying salesperson. Route the observation to a product or commercial owner who can check fit, value and customer context. Close the loop so the technician can see what happened.

A study of 127,000 B2B service visits at DMG Mori found that technicians offered less often in some of the service situations with the highest sales success. The useful lesson is operational: capture the service situation, qualify it consistently and support the field with a clear next step instead of relying on individual selling instinct.

Run fleet campaigns with machine-level control

Some triggers affect a population: a safety recall, software update, known component fault, end-of-life notice or standardized upgrade. Define the campaign scope from product, component, serial range, configuration, geography and service state. Freeze the rule version used to select the population.

Create an item for every affected asset. Track contact, inspection, parts, work, completion, refusal and unreachable status separately. Update the population when ownership or asset status changes. Maintain an exception path for machines with field modifications or incomplete configuration.

Salesforce’s product service campaign workflow covers recalls, upgrades, maintenance, feedback, warranty upsell and customer communication. That breadth shows why the campaign record needs a precise purpose and completion rule. “Contacted” does not complete a recall; “work executed on the affected asset” does.

Connect the offer to quote and execution

Carry the asset, trigger, configuration, proposed work, price basis and promised result into the quote. When the customer accepts, reconcile the PO or authorization with the approved offer and create the correct parts order, repair order, project or field-service work order.

Execution needs the same identity. The technician or depot should see the affected asset, current configuration, work instruction, parts, tools, evidence and completion criteria. If delivery uncovers a different machine state, stop, update the scope and obtain approval before proceeding.

Microsoft’s Field Service overview describes work orders arising from service cases, agreements, portals and IoT data, with used parts and completed work updating service history and customer assets. The pipeline closes only after the physical work and asset record agree.

Update the installed state after delivery

Record what changed: removed and installed parts, serial numbers, software, configuration, warranty, service interval, meter, documents, test result and next due event. Link the completion evidence to the asset and preserve the effective date.

This update prevents the next campaign from recommending completed work. It also changes spare-parts fit, service instructions, entitlement and future upgrade eligibility. If a distributor or partner performed the work, require the same closeout data before campaign completion.

Sales outcome alone cannot close the loop. A won quote that never ships should remain unfulfilled. A shipped kit that the customer never installs should not change the as-maintained state. Track commercial, delivery and installation milestones separately.

Measure the whole conversion chain

Revenue tells you what reached the end. Measure the loss at each earlier step so the team can improve the system. Segment results by trigger, offer, product family, asset age, region, channel and evidence quality.

Worked example: controller obsolescence across 286 machines

An industrial packaging OEM has 286 active machines with a D4 control family. Engineering confirms that one processor and one display will lose supplier support in eighteen months. The first population rule finds 173 machines in the relevant serial range. Configuration records show that 41 already received the new controller, leaving 132 candidates.

The team checks current ownership, operating status and installed revision. Eleven machines have retired, nine moved to customers served by a distributor and eighteen need a photo or controller export. Ninety-four assets pass the first technical gate. Product engineering releases two kits because early and late wiring differ.

Commercial qualification uses site criticality, current spares, recent failures, planned shutdowns and buyer access. Fifty-eight assets across twenty-three sites enter the qualified pipeline. The OEM groups multiple machines at each site, gives the distributor its nine eligible items and suppresses twelve low-duty machines whose customers hold adequate spares.

The team presents proposals for forty-six machines during scheduled account reviews. Customers accept thirty-four upgrades, defer eight to the next shutdown and decline four. Execution completes thirty machines in the quarter. Four wait for customer shutdown dates. Each completed work order records the new controller, wiring kit, software and warranty. The next campaign query excludes those machines automatically.

Assign ownership across the chain

Installed-base operations should own population quality and asset identity. Product or service engineering should own technical applicability. The aftermarket product owner should own the offer, price logic and readiness. The account or channel owner should own customer access and decision. Service or order operations should own delivery. One pipeline owner should monitor handoffs and aging across the full chain.

Give every rejection and delay a destination. Missing configuration returns to asset-data work. An unreleased kit returns to product engineering. No customer contact returns to account ownership. No capacity returns to service planning. This avoids a large shared queue with no responsible team.

Run a short weekly review around exceptions: new high-impact triggers, old qualification work, offers blocked by supply, customer decisions due, accepted work without capacity and completed work missing asset updates.

Use the systems that own each fact

The installed-base record should own asset identity, location and configuration history. PLM and engineering systems should own applicability, released kits and service documents. Field service should own observations, work orders and completion. ERP should own parts, price, order, inventory, invoice and cost. Contract systems should own entitlement and renewal. CRM should own contacts, opportunities and customer actions.

The pipeline application should reference those records with stable IDs and dates. It should calculate candidate populations, apply rules, route work and show the evidence without copying every source object into a new master.

Manufacturing Cloud events and milestones provides one example of capturing asset events such as delivery, extended warranty, service due and recall, then initiating timely action. The useful pattern is event to asset to controlled action. The source system remains accountable for the underlying fact.

How Bourne turns fleet evidence into work

Bourne reads assets, configuration, service events, contracts, parts demand, engineering notices, telemetry, customer records and commercial history from the systems that hold them. It applies approved trigger and eligibility rules, groups duplicate signals and prepares an asset-specific candidate with the source evidence visible.

The application routes technical checks to engineering, account and channel checks to commercial owners, and offer gaps to the aftermarket product team. It drafts the machine list, customer reason, recommended scope and next action. People approve technical claims, customer outreach, price and commitments.

After acceptance, Bourne carries the approved scope into the quote, order or work system. It checks delivery milestones and writes the completed configuration and next service event back to the installed-base record. The result improves later campaigns, parts planning and account reviews.

Pilot one offer on one verified fleet

Choose a repeatable offer with a clear asset rule and customer benefit: a controller retrofit, service-agreement renewal, overhaul interval, wear-parts plan or safety update. Limit the first population to one product family, region or service team. Confirm active status and configuration before generating candidates.

Define the trigger, eligibility rule, evidence threshold, stage gates, owner, suppression and closeout data. Review every candidate manually during the pilot. Track why each item qualified, stalled, converted or failed. Connect accepted work to execution and update the installed state.

The pilot passes when the team can reproduce the population, explain every recommendation, avoid duplicate customer contact, forecast from real stage evidence and prove that delivered work changed the correct assets. Then add the next trigger or fleet.

Arda Bulut

Arda Bulut is the co-founder and CTO of Bourne and HockeyStack. He leads engineering at Bourne, building the platform people use to create AI products, agents and automations.