How Do I Track Attach Rate and Add-On Sales by Rep?

What per-rep attach tracking replaces

Most teams already track attach in some form, and almost all of them track it in a way that cannot be coached. Understanding what you are replacing matters more than picking a tool, because the tool choice follows from which of these failure modes you currently have.

The store or team average. The most common setup: one number on a monthly report showing the location attached warranties to 38% of transactions. It is a real number and it is completely useless for management. A floor with eight reps can hit 38% with three reps at 70% and five reps at 15%, or with all eight sitting flatly at 38%. Those are entirely different problems — one is a coaching problem with three names on it, the other is a process or product problem — and the average is identical in both cases. Every hour spent debating the store number is an hour not spent watching the actual distribution.

Total add-on revenue by rep. A step better, and still misleading. Revenue-by-rep conflates volume with behavior. Your highest-volume rep will usually top the add-on revenue chart even with the worst attach discipline on the floor, simply because they touch more transactions. If Rep A writes 200 tickets and attaches on 25 of them, and Rep B writes 60 tickets and attaches on 30, Rep A shows more add-on dollars and Rep B is twice the bundler. Attach rate — a ratio, not a total — is what normalizes for volume. Track both, but rank on the rate.

The commission statement as a proxy. Some teams assume that because add-ons pay a spiff, the number takes care of itself. It does not. A spiff tells a rep that add-ons are worth money; it does not tell them where they stand relative to peers, and it does not surface the rep who has quietly written zero warranties in six weeks. Comp is a lagging incentive. A published per-rep rate is a leading one, and the two work best stacked.

Manager gut feel. "Marcus is good at add-ons" is a real signal and a bad system. It survives one manager and dies at the next, it cannot be audited, and it is systematically biased toward the loud reps. Write it down or it does not exist.

Ad-hoc SQL when someone asks. The report that only exists when the VP requests it is a report nobody manages to. Attach rate has to appear on the same cadence as the sales number — weekly minimum, daily on a retail floor — or reps treat it as a compliance artifact rather than part of the job.

The replacement in every case is the same shape: attach rate as a standing, named, per-rep line with a target next to it, refreshed on a schedule, visible to the rep whose name is on it. Everything else in this page is mechanics for getting there.

How to choose your tracking path

There are really only three architectures, and the right one falls out of where your line-item data already lives and how many people need to see the result.

Path one: the spreadsheet scorecard. Export deal or transaction line items, pivot by rep, categorize SKUs, done. This is the correct starting point for teams under roughly 15 reps or single-location retail. It costs nothing but an analyst's afternoon, it is fully transparent, and it forces you to make the SKU categorization decisions explicitly rather than inheriting someone's defaults. The failure mode is staleness — the sheet is only as current as the last person who remembered to refresh it — so pair it with a standing calendar block, not good intentions. Many teams run a spreadsheet for two quarters, learn what the real definitions need to be, then port the now-proven logic into a system.

Path two: native CRM or POS reporting. If your line items already live in Salesforce, HubSpot, a retail POS, or a field-service platform like ServiceTitan, build the report where the data is. You get automatic refresh, no export step, and attach rate sitting next to pipeline where reps already look. The work is upfront: you need a clean product catalog with a category field, and you need deal ownership to be accurate. In practice the catalog hygiene is the whole project — most orgs discover that the same warranty exists as four differently-named SKUs, which is why their historical attach numbers never reconciled.

Path three: a dedicated scorecard, gamification, or BI layer. Tools that sit on top of the CRM and handle weighting, leaderboards, TV displays, or multi-source blending. Worth it when you have multiple locations, multiple data sources, or a floor culture that responds to visible competition. Not worth it before you have a definition you trust, because a leaderboard broadcasting a bad metric teaches the whole team to game a bad metric.

The decision below is not really about tools. It is about how many data sources you have and whether the number has to be self-serve for reps.

A note on sequencing that saves rework: whichever path you pick, write the definition document before you build anything. One page listing what counts as core, what counts as add-on, what happens to bundles sold as a single SKU, how returns are handled, and what the denominator is. Two people building from that page will produce the same number. Two people building from a Slack thread will not.

What it costs and what it moves

Budget realistically and the project survives contact with the quarter. Here is what the work actually consumes.

Time to first credible number. For a single-source spreadsheet: half a day to a day, most of it spent on SKU categorization rather than formulas. For a native CRM build: one to three weeks, and the variance is entirely product-catalog cleanliness. If your catalog has consistent category fields you are at the low end; if warranties and service plans are entered as free-text line items or miscellaneous SKUs, you are doing a cleanup project first and should scope it as one. For a multi-source BI build: four to eight weeks including data modeling and reconciliation against the source systems.

Ongoing maintenance. Budget two to four hours a month for SKU categorization drift — new products launch, old ones retire, someone creates a duplicate SKU. This is real and it is the reason attach dashboards silently rot. Assign it to a named person in RevOps, not to "the team."

Tooling cost. The spreadsheet is free. CRM reporting is included in seats you already pay for. Dedicated scorecard, gamification, and BI layers span a wide range — typically single-digit to low-double-digit dollars per user per month at the lighter end, with enterprise platforms quoted rather than listed. Do not buy at this layer until the spreadsheet version has run long enough to prove the definition, because the cheapest way to discover your denominator is wrong is to discover it in a free file.

What actually moves. Be honest about the mechanism: publishing a number does not sell warranties. What it does is convert an invisible behavior into a visible one, which makes three things possible that were not possible before. First, targeted coaching — you know which four reps to sit with instead of running a floor-wide refresher that the good reps sit through resentfully. Second, hiring and ramp signal — a new rep's attach rate at week six tells you whether the onboarding covered the bundle conversation, and it tells you months before their revenue number would. Third, diagnosis of non-rep causes — when every rep's attach rate drops in the same week, the problem is not the reps, it is a price change, a stockout, a systems change, or a competitor. That last one is underrated: a per-rep view is also the fastest way to prove something is *not* a rep problem.

Realistic expectations on the number itself. Attach rate responds to attention, and the first movement is usually fast and partly artificial — reps start recording add-ons they were already selling but not logging. Treat the first three to four weeks after launch as a data-hygiene correction, not a performance gain, and set your real baseline after it stabilizes. Sustained improvement after that comes from coaching and product mix, and it is incremental. Anyone promising you a step-change from a dashboard is selling a dashboard.

Where it can go wrong. Two failure modes worth pricing in. Discounting drift: if reps are measured on attach and not on margin, some will attach by discounting the add-on to near-zero. Track attached add-on average selling price alongside the rate. Customer-fit drift: pushing attach hard on a floor without a matching quality metric produces returns and cancellations. Watch add-on return and cancellation rate as a paired guardrail. A rate that climbs while cancellations climb faster is not a win, and RevOps should be the function that says so out loud.

Building it and handing it off

The build has a natural order, and skipping steps here is why so many attach dashboards get built twice.

Define the categories. Every SKU in the catalog gets exactly one of three labels: core, add-on, or excluded. Core is the anchor product. Add-on is the accessory, warranty, service plan, extended coverage, install, training, or upsell module. Excluded covers shipping, fees, taxes, adjustments, and anything that would inflate the denominator without representing a selling decision. The excluded bucket is the one people forget and it is the one that quietly breaks the number.

Handle the edge cases explicitly. Bundles sold as one SKU — decide whether that counts as an attach or gets its own category, and be consistent. Returns and cancellations — decide whether attach rate is measured on booked or on net-of-return, and prefer net if your return window is short enough to make it practical. Split credit — if two reps touch a deal, decide whether attach credits the closer, splits, or credits both; splitting is more accurate and harder to coach against, so most floors credit the closer and accept the noise. Multi-line deals — a deal with three add-ons is still one attached deal for rate purposes; use a separate "add-ons per deal" metric if depth matters to you.

Pick the denominator and never change it quietly. Attached deals divided by total deals is the common definition and the one most people mean. Attached units divided by total units is the retail variant and it behaves differently on multi-unit tickets. Both are defensible; mixing them across periods is not. If you do change it, restate history or you will spend a quarter arguing about a number that never moved.

Build the pull. Line items, not deal headers — deal-header reporting cannot see attach at all, and this is the single most common reason a CRM "already reports attach" and the number is wrong. Join to rep via deal owner or transaction operator, group by rep and period, and compute the rate.

Set targets by segment, not globally. A rep selling into enterprise with a six-month cycle and a rep on a retail floor should not share an attach target. Segment by product line, deal size band, or channel, and set the target from your own trailing distribution — typically the 60th to 75th percentile of current performance, which is reachable enough that the bottom half engages and demanding enough that the top half does not coast.

Hand it off properly or it dies. The handoff is where these projects fail, not the build. Three things have to transfer: ownership of the SKU categorization (named person, monthly recurring task), ownership of the refresh (automated where possible, calendared where not), and ownership of the conversation (the frontline manager, in the existing 1:1, not a separate meeting). If attach rate gets its own meeting it will be cancelled within two months. If it becomes a standing line in the 1:1 the manager already runs, it survives.

Instrument the coaching, not just the number. The rate tells you who; it does not tell you why. A rep at 12% might never raise the add-on, might raise it and get objected out, or might work a segment where the add-on genuinely does not fit. Those need three different coaching responses. Conversation-intelligence tooling can surface whether the pitch is even happening; absent that, a manager listening to four calls or shadowing two hours on the floor answers it in an afternoon. Do that before you build a training program for a problem you have not diagnosed.

Adjacent metrics this unlocks

Once line items are tagged and attributed by rep, you have built the substrate for a set of neighboring measures that are considerably harder to stand up from scratch — and most teams get more value from these than from the attach number that justified the work.

Add-on average selling price by rep. The margin guardrail mentioned earlier. Reveals the rep who hits target by giving the warranty away, and the rep with a lower rate but a much richer mix. Together, rate and ASP give you attached revenue per deal, which is the number a CFO actually cares about.

Attach by product line. Rotate the same data and you learn that attach on one category runs 55% while another sits at 9%. That is rarely a rep problem — it is usually pricing, packaging, a missing pitch, or a product nobody believes in. This is where RevOps earns its keep, because it is the finding that redirects effort away from coaching and toward a fixable upstream cause.

Ramp curves for new hires. Plot attach rate by weeks-since-start across your last two years of hires. You get an expected curve, and any new rep tracking below it is an onboarding intervention available months earlier than revenue would have flagged it.

Renewal and retention correlation. In subscription and service businesses, join attach at the point of sale to renewal outcomes twelve months later. Many teams find that accounts with a service plan or onboarding package attached renew at a meaningfully higher rate — and if that holds in your data, attach stops being a margin play and becomes a retention play, which changes how you weight it and how you fund it.

Cross-functional reuse. The tagged line-item table feeds finance's margin analysis, marketing's attach-by-campaign view, and support's "which configurations generate tickets" question. Build it once as a governed dataset rather than a dashboard-specific query, and three other functions stop maintaining their own conflicting versions of the same categorization.

Comparable patterns outside sales. The structure generalizes: service organizations measure attach as the percentage of jobs where a technician quotes a needed repair, healthcare-adjacent retail measures it as protection-plan capture, and subscription businesses measure it as expansion-module rate at initial close. Same mechanic every time — a per-person ratio, a named target, a visible publication, a coaching conversation. If you have built it once, you can build it for the next function in a week.

Related questions

Should attach rate be part of commission?

Usually yes, as a modifier or accelerator rather than the primary driver. Paying a separate accelerator on add-on revenue makes bundling non-optional without distorting the core motion. Pair it with a margin or cancellation guardrail so reps cannot earn it by discounting.

How often should I publish per-rep attach rate?

Weekly for most teams, daily on a high-volume retail floor. Monthly is too slow to coach against — a rep gets four weeks of reinforced bad habit before anyone says anything. Match the cadence to how quickly the behavior can change.

What is a good attach rate?

There is no universal benchmark; it depends entirely on product mix, price point, and channel. Set your target from your own trailing distribution — commonly the 60th to 75th percentile of current rep performance — and move it as the floor improves.

Can I track attach rate without a CRM?

Yes, if your POS or order system exports line-item detail with an operator or salesperson field. That is the only real requirement. Deal-header-only exports cannot produce attach rate at any tool tier, so check for line-item granularity before scoping anything.

Who should own the attach metric?

RevOps owns the definition, the data pipeline, and the SKU categorization. The frontline sales manager owns the conversation and the coaching. Splitting it any other way produces either an accurate number nobody uses or a used number nobody trusts.

FAQ

What exactly counts as an attach?

An attach is any add-on line sold alongside the core product on the same transaction — accessory, extended warranty, service plan, install, training, or an upsell module. The core unit is the anchor; everything bundled onto it counts. Fees, shipping, and tax adjustments should be excluded so they do not inflate the denominator. Write the classification down once and apply it consistently, because the most common cause of an attach number nobody trusts is two people using two definitions.

Why is attach rate better than total add-on revenue?

Revenue totals scale with volume, so your busiest rep tops the chart regardless of bundling discipline. A rate normalizes for how many transactions each rep touched, which is what makes it comparable across the team and coachable in a 1:1. Track add-on revenue too — it is what the business banks — but rank and coach on the rate.

My CRM says it already reports attach. Why is the number wrong?

Almost always because the report is built on deal headers rather than line items. Header-level data knows a deal closed and for how much; it cannot see which SKUs were on it. The second most common cause is a dirty product catalog where the same warranty exists under several SKU names, so a portion of real attaches never get counted.

How do I handle deals where two reps were involved?

Pick one rule and publish it. Most floors credit the closing rep and accept some noise, because split credit is more accurate but much harder to coach against — you cannot tell a rep their rate is low when half their deals carry partial credit. If splits matter to your comp plan, report both a credited rate and a touched rate, and coach on the credited one.

Won't measuring attach push reps to sell things customers don't need?

It can, which is why the rate never ships alone. Pair it with add-on return and cancellation rate, and with attached add-on average selling price. A rate that climbs while cancellations climb faster is a net loss, and the paired metric is what lets RevOps catch it in the same reporting cycle rather than two quarters later in a churn review.

How long before the number is trustworthy?

Expect three to four weeks of noise after launch, driven by reps logging add-ons they were already selling but not recording. Set your real baseline after that settles. Then give it a full quarter before drawing conclusions about individual reps, since attach rate on low-volume territories is genuinely noisy at small sample sizes.

Sources

• Salesforce — reports and dashboards documentation: https://help.salesforce.com/s/articleView?id=sf.reports_dashboards.htm
• HubSpot — products and line items in the CRM: https://knowledge.hubspot.com/products/use-products
• Microsoft Power BI — documentation and data modeling guidance: https://learn.microsoft.com/en-us/power-bi/
• Tableau — help and dashboard building documentation: https://help.tableau.com/current/pro/desktop/en-us/default.htm
• Gong — revenue intelligence platform: https://www.gong.io/
• QuotaPath — commission and quota attainment tracking: https://www.quotapath.com/
• Harvard Business Review — research and analysis on sales performance management: https://hbr.org/topic/subject/sales
• McKinsey & Company — growth, marketing, and sales insights: https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights
• Google Sheets — pivot table documentation: https://support.google.com/docs/answer/1272900

Related on PULSE

• [How Do I Know Where, When, and How Many People to Schedule at Each of My Multi-Unit Retail Locations?](/knowledge/tl0001)
• [How Do I Figure Out How Many People to Schedule Each Day and at What Times for My Single Store?](/knowledge/tl0002)
• [How Do I Know How Many Cooks and Servers to Schedule Each Shift at My Pizza Restaurant?](/knowledge/tl0003)
• [How Many Salespeople Should I Schedule Each Day on My Furniture Store Floor?](/knowledge/tl0004)
• [How Do I Decide How Many Reps to Schedule at Each Store in My Mattress Retail Chain?](/knowledge/tl0005)