How Do I Score My Reps on New Logo Versus Expansion in 2026?

The outcome you should expect

The practical outcome of splitting the scorecard is that you stop being surprised. Before the split, a rep who books $600K looks identical to another rep who books $600K, and you have no idea that the first one landed nine new accounts while the second one re-signed a single enterprise renewal with a 12 percent price uplift and never ran a new discovery call all quarter. One of those reps is building next year's base and the other is harvesting it. A single number hides that difference completely, and it usually hides it for two or three quarters — right up until new-logo pipeline dries out and the number that used to look fine collapses with no warning.

With two weighted lines, the imbalance shows up in week three, not month nine. You get a per-rep picture that reads something like: new-logo bookings level 2, expansion bookings level 5, gross retention level 4, pipeline created level 1, win rate level 3. That rep is a farmer. Nothing wrong with a farmer — but now you know, the rep knows, and the coaching conversation has a specific subject instead of a vague "you need more activity." The visible gap between level 2 and level 5 on the two revenue lines is the whole point of the exercise.

The second outcome is that territory and account assignment gets rational. Once you can see which reps score high on new logo and which score high on expansion, you can stop pretending every seat is interchangeable. Some organizations respond by specializing — a net-new team and an account-management team with different scorecards and different weights. Others keep the hybrid seat but rebalance the book so a rep with a 40-account installed base isn't also expected to prospect cold. Either decision is defensible. What is not defensible is running hybrid seats while measuring only the blended total, because that combination reliably produces reps who quietly optimize for whichever motion is easier that quarter, which is almost always expansion.

The third outcome is forecast quality. New-logo deals and expansion deals have different cycle lengths, different win rates, and different discounting behavior. Expansion into a happy account might close in 30 days at an 65 percent win rate; a competitive net-new displacement might take 140 days at a 22 percent win rate. If those two flows are blended into one pipeline number, your coverage ratio is meaningless, because you are applying one average conversion rate to two populations that behave nothing alike. Splitting the scorecard forces you to split the pipeline reporting, and split pipeline reporting is what makes a coverage ratio worth quoting to a board.

Expect one uncomfortable outcome too: some of your top-performing reps will score worse than they did under the blended number. That is the system working, not the system failing. Plan for the conversation before you publish the first matrix.

What drives that outcome

The mechanism is arithmetic, and it is worth writing out explicitly because the whole system rests on it. Composite score equals the sum of (weight × level) across every KPI on the card. Weights are percentages that total 100. Levels are integers from 1 to 5, where 3 is "meets the bar," 5 is "top of the team," and 1 is "materially behind." Nothing more complicated than that — the sophistication lives in which lines you choose and how you weight them, not in the math.

A workable starting card for a hybrid seat at a growth-stage company looks like this. New-logo bookings at 30 percent. Expansion and upsell bookings at 25 percent. Gross retention on the assigned book at 15 percent. New-logo pipeline created at 15 percent. Win rate at 10 percent. Multi-threading or activity quality at 5 percent. A rep at levels 2 / 5 / 4 / 1 / 3 / 3 scores (0.30×2) + (0.25×5) + (0.15×4) + (0.15×1) + (0.10×3) + (0.05×3) = 0.60 + 1.25 + 0.60 + 0.15 + 0.30 + 0.15 = 3.05. A rep at 5 / 2 / 3 / 4 / 3 / 4 scores 1.50 + 0.50 + 0.45 + 0.60 + 0.30 + 0.20 = 3.55. Both reps might have booked identical total dollars. The composite says the second one is closer to complete.

Three design choices drive whether that arithmetic actually changes behavior.

First, the level definitions have to be written down as concrete thresholds, not adjectives. "Level 3 on new-logo bookings = $300K–$449K in the trailing four quarters" is a scorecard. "Level 3 = solid performance" is a Rorschach test that produces a different score depending on who fills it in. Write the bands, publish them, and hold them for at least two full quarters before you move them.

Second, the definition of what counts as new logo versus expansion has to be settled before anyone is scored, because the boundary cases are where the credibility is won or lost. A dormant account that churned 18 months ago and comes back — new logo or expansion? A new business unit inside an existing parent with its own budget and its own procurement — new logo or expansion? A subsidiary in a different country? The answer matters less than writing it down. Most teams land on something like: any contract with a new billing entity that has had no active subscription for 12 or more months counts as new logo; anything sold into an entity with an active contract counts as expansion; new divisions of an existing parent count as expansion unless they buy through a separate paper contract, in which case they count as new logo. Publish whatever rule you pick in the same document as the matrix.

Third, the weights have to be set by the same people who set the growth plan. If the board asked for market share and the finance plan assumes 60 percent of net-new ARR comes from new logos, then a scorecard weighting expansion at 40 percent and new logo at 15 percent is actively fighting the plan. RevOps should build the card, but the weights are a leadership decision, and they should be revisited on the same cadence as the plan — usually annually with a mid-year check, not every month.

Benchmarks and realistic ranges

Be careful with external benchmarks here, because the two motions are measured in different units and it is easy to quote a number that does not mean what you think it means.

On the expansion side, the standard measure is net revenue retention — the revenue from a cohort of existing customers a year later, including upsell and price increases, minus downgrades and churn. Healthy SaaS businesses generally run NRR above 100 percent; strong enterprise-focused companies commonly land somewhere in the 110 to 125 percent range, and best-in-class outliers go higher. Below 100 percent means the installed base is shrinking on its own and expansion is not covering churn and downgrades. Gross retention, which excludes upsell, sits lower by definition — mid-80s to mid-90s percent is a common range depending on segment, with SMB at the low end and enterprise at the high end. Do not benchmark a rep against a company-level NRR figure directly; a single rep's book is too small a sample for the number to be stable quarter to quarter, which is exactly why the scorecard converts it to a 1-to-5 level against your own internal distribution rather than against an industry average.

On the new-logo side there is no clean universal benchmark, because logo counts depend entirely on average deal size. A rep selling $15K annual contracts might need 30 new logos a year; a rep selling $400K contracts might need three. So set the bands off your own trailing data. The practical method: pull the last four to eight quarters of closed-won by rep, split by motion, and look at the distribution. Set level 3 at the median, level 4 at roughly the 70th–75th percentile, level 5 at roughly the 90th, level 2 at roughly the 30th, and level 1 below that. That gives you a card calibrated to reality rather than to hope, and it means roughly a third of the team lands at level 3 on any given line — which is what "meets the bar" should look like.

For the weights themselves, a few patterns hold up across most organizations. An early-stage company still proving the market and chasing logo count typically weights new logo somewhere in the 35 to 45 percent range and expansion at 15 to 20 percent. A scaled company protecting a large base flips it: expansion and gross retention together in the 40 to 50 percent range, new logo at 20 to 25 percent. A company in a deliberate efficiency year — protecting margin rather than chasing growth — pushes gross retention up to 20 to 25 percent on its own, which is unusually high but appropriate when churn is the presenting problem. Pipeline created almost always deserves 10 to 20 percent regardless of stage, because it is the only leading indicator on the card and it is the line that catches a rep coasting on the base before the bookings gap appears.

Two more calibration notes. Keep the total number of scored lines between five and seven. Below five and you have not actually separated the motions; above seven and each weight gets so small that moving a level barely changes the composite, which kills the incentive. And keep the smallest weight at 5 percent or above — a 2 percent line is decoration, and reps correctly ignore decoration.

Risks, edge cases, and failure modes

The most common failure is that the scorecard gets built and never gets published. RevOps produces a beautiful weighted matrix, leadership reviews it in a QBR, and no rep ever sees their own levels. A scorecard nobody can see is a reporting artifact, not a management system. Reps change behavior in response to a number they can watch move; they do not change behavior in response to a number that appears once a quarter in a slide they hear about secondhand. Publish per-rep levels on a cadence — weekly or biweekly is typical — and make the bands visible so anyone can compute their own score.

The second failure is gaming the classification boundary. Once new-logo credit is worth more than expansion credit, reps get creative about what counts as a new logo. A new department at an existing customer gets papered as a separate contract. A dormant account that was going to renew anyway gets held past the 12-month mark so it re-enters as net-new. This is not a character problem, it is a design problem: any credit boundary with a value gap across it will be tested. Mitigations that work: require a deal-desk or RevOps sign-off on the motion classification for any deal above a dollar threshold, audit a sample of classifications each quarter, and keep the credit gap between the two motions moderate rather than extreme. If new logo pays 3x expansion, expect aggressive reclassification; if it pays 1.4x, the incentive to game it mostly disappears.

Third failure: unfair books. A rep handed 12 greenfield territories and a rep handed 45 mature accounts cannot be scored on the same weights and both be treated justly. The fix is not to abandon the two-line card — it is to acknowledge that book composition is an input. Two workable approaches. Either run two scorecard profiles (hunter profile weighted toward new logo, farmer profile weighted toward expansion and retention) and assign reps to a profile with the assignment reviewed annually, or keep one card and normalize the level bands per segment so that "level 3 on new logo" means something different for an enterprise seat than for a mid-market seat. Either works. What does not work is one card, one set of bands, and wildly uneven books.

Fourth: double-counting when a CSM or account manager is also on the expansion line. If the AE and the CSM both get full expansion credit on the same deal, the company is scoring 200 percent of the revenue it actually booked, and the scorecard stops reconciling to finance. Decide the split explicitly — full credit to both is acceptable if it is a *scorecard* and not a *comp* mechanism, but if it drives pay, split it, or give the CSM a retention-weighted card and the AE an expansion-weighted one so the same dollar is not paid twice at full rate.

Fifth: weight thrash. The whole appeal of a weighted matrix is that you can re-weight when strategy shifts. The risk is that leadership discovers this and starts re-weighting monthly. Reps need at least a quarter, preferably two, to change behavior in response to a weight change — sales cycles are longer than the reporting cycle. Re-weighting more often than that produces reps who stop believing the card and just keep doing whatever they were doing. Set a rule: weights change at plan boundaries, not on impulse, with a documented reason.

Sixth, and subtlest: the composite can mask a level 1. A rep scoring 5s on four lines and a 1 on new logo can still post a respectable composite. If new-logo pipeline is genuinely a must-have, add a floor rule — no rep is rated above "meets expectations" overall with any revenue line at level 1 — rather than trying to fix it by cranking the weight. Floors handle non-negotiables more honestly than weights do.

A practical rollout plan

Run it in four phases over roughly a quarter. Rushing to a scored, paid card in three weeks is the reliable way to get it rejected by the field.

Phase one — define, one to two weeks. Settle the motion classification rule (new logo versus expansion, including the dormant-account and new-division edge cases) and get it signed off by sales leadership, finance, and RevOps in the same room. Then confirm the CRM can actually produce the split. This is usually where the real work is: most CRMs have an opportunity type field that has been filled in inconsistently for years. Audit the last four quarters of closed-won, fix the historical classification, and add validation so the field cannot be left blank or free-typed going forward. If you cannot produce a clean trailing-eight-quarter split by rep and motion, you are not ready to score anyone.

Phase two — calibrate, one to two weeks. Build the card: five to seven lines, weights summing to 100, agreed with leadership against the current plan. Set the level bands off the trailing distribution you just cleaned up. Then run the card retroactively against the last two closed quarters and look at the output. Two sanity checks: does the ranking pass the smell test with the sales leaders who know these reps, and is the spread wide enough to be meaningful? If everyone lands between 3.1 and 3.4, the bands are too generous and the card will not discriminate. Adjust the bands, not the reps.

Phase three — publish and shadow, four to six weeks. Show every rep their own card and the full band definitions, and explicitly say this quarter is informational — it does not touch pay. Hold one manager-led conversation per rep walking through their levels and naming the one line to move. Collect the objections; the good ones are almost always about book composition or classification edge cases you missed, and both are worth fixing before the card has teeth. Expect to revise something.

Phase four — wire the teeth, next full quarter. Attach consequences. That can mean comp — separate accelerators or rates on the new-logo and expansion components — or it can mean territory review, promotion criteria, and coaching priority. Comp is the strongest lever and the hardest to reverse, so if you are unsure, start with non-comp teeth for a quarter and add the pay linkage at the next plan boundary. Whatever you attach, keep publishing levels on the same cadence so a rep can watch a line move.

Then hold it. Review weights at the annual plan, review bands after two quarters of data, and audit classification quarterly. The card only compounds if it stays stable long enough for reps to trust it.

Related questions

Should new logo and expansion sit on the same quota or separate quotas?

Separate components under one total is the cleaner design. Give each rep a new-logo number and an expansion number that sum to their total, so the scorecard lines have a matching quota line. One blended quota lets reps fill it with whichever motion is easier.

How do I score a rep whose territory has almost no installed base?

Assign them a hunter scorecard profile with new logo weighted at 40 to 45 percent and expansion dropped to 10 percent or removed. Uneven books are a weighting problem, not a scoring problem. Review profile assignment annually when territories are redrawn.

Does expansion credit belong to the AE or the CSM?

Depends on who runs the deal. If the CSM sources and closes, credit them; if they source and hand off, split it. For pay, split explicitly so one dollar is not funded twice. For a non-comp scorecard, dual full credit is acceptable.

How often should the weights change?

At plan boundaries — annually, with a mid-year check if the growth mix genuinely shifts. Reps need a quarter or two to change behavior, so monthly re-weighting produces noise instead of movement, and it teaches the team to ignore the card.

FAQ

What exactly is the composite score formula?

Composite equals the sum of (weight × level) across every scored line. Weights are percentages totaling 100; levels are integers 1 through 5 assigned from published thresholds. With six lines the composite lands between 1.00 and 5.00. Keep the math this simple — the value is in the line selection and the weighting, not in a more elaborate formula.

How do I decide the split between new logo and expansion weight?

Match it to the finance plan. If the plan assumes most net-new ARR comes from new logos, weight new logo higher — commonly 35 to 45 percent early-stage, with expansion at 15 to 20. A scaled company protecting a large base inverts that, putting expansion and gross retention together at 40 to 50 percent. Leadership sets the weights; RevOps builds the card.

What counts as a new logo versus expansion?

Write the rule down before scoring anyone. A common standard: a new billing entity, or an account with no active subscription for 12 or more months, counts as new logo; anything sold into an entity with a live contract counts as expansion. New divisions of an existing parent usually count as expansion unless they contract separately. The exact rule matters less than publishing it and auditing against it.

What if a rep is excellent at expansion and weak at new logo?

The composite will reflect it, and that is the point — the imbalance becomes a specific coaching subject instead of a vague concern. If new-logo production is truly non-negotiable, add a floor rule preventing an overall "meets expectations" rating with any revenue line at level 1, rather than distorting the weights to force the outcome.

Can reps game the classification to grab new-logo credit?

Yes, if the credit gap is large. Mitigate it three ways: require deal-desk sign-off on motion classification above a dollar threshold, audit a sample of classifications quarterly, and keep the credit differential moderate. A 1.4x differential rarely gets gamed; a 3x differential reliably does.

Do I need software to run this?

No. A spreadsheet with published bands and a weighted-sum formula runs the method fine, and many teams start there. The cost is upkeep and the risk of a stale sheet nobody refreshes. CRM dashboards, sales-performance platforms, and incentive-comp tools automate the rollup once the definitions and weights are settled — but settle those first, because no tool will decide them for you.

Sources

• https://www.saastr.com/net-revenue-retention-what-is-good-what-is-great/
• https://openviewpartners.com/blog/net-revenue-retention/
• https://www.bvp.com/atlas/state-of-the-cloud
• https://www.forentrepreneurs.com/saas-metrics-2/
• https://hbr.org/2015/04/motivating-salespeople-what-really-works
• https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/the-new-b2b-growth-equation
• https://www.gartner.com/en/sales/topics/sales-performance-management
• https://help.salesforce.com/s/articleView?id=sf.reports_dashboards.htm&type=5

Related on PULSE

• How Do I Score My CSMs on Retention and Expansion?
• How Do I Score My Reps on Partner-Sourced Pipeline?
• How do you separate NRR, GRR, and logo retention when board auditors ask which is 'real'?
• How do you compensate a sales rep who lands a strategic-but-low-ARR logo (e.g. brand-name reference customer)?
• How should a 2027 sales org structure ramped pricing for new logo expansion?