Blog · How-to guides · Sales teams
Compare customer contribution, purchase cadence and acquisition cohorts before choosing a ranking method. A worked B2B example shows why different questions produce different account priorities.
RFM analysis alternatives are most useful when the business decision depends on contribution, expected reorder timing or customer age. Choose the measure that answers the question, then keep the supporting evidence visible. A customer can be first in revenue, third in contribution and first on the call list without any contradiction.
RFM summarizes recency, frequency and monetary value. The existing RFM analysis for B2B customers guide owns the scoring method and its B2B adjustments; the choice here is whether a different measure should lead the decision.
| Business question | Approach | Files needed | Useful output |
|---|---|---|---|
| Which accounts contribute most after variable costs? | Customer contribution | Net sales and attributable variable costs | Contribution dollars and rate by account |
| Which repeat buyers need a conversation? | Purchase cadence | Customer IDs, distinct orders and dates | Days since purchase relative to expected interval |
| Are newly acquired customers coming back? | Acquisition cohorts | First purchase and subsequent order dates | Repeat purchase rate at equal customer ages |
| Which existing accounts have room to grow? | Category coverage | Customer segments, eligible categories and purchases | Unbought categories supported by account fit |
These approaches can sit beside one another. A sales manager needs an account contact order; finance needs the economics of serving those accounts. An acquisition manager needs evidence about groups of new customers, which is a different level of analysis.
Do not create a weighted score before agreeing on the action. Combining urgency and contribution can hide a large delayed account behind an otherwise healthy total. A separate flag preserves the reason for review.
A US industrial distributor reviews four accounts for the twelve months ending September 30, 2026. Net revenue already includes credits and discounts. Variable costs include product costs and the variable fulfillment and selling costs attributed to the same sales; fixed office costs are excluded.
| Account | Net revenue (USD) | Variable costs (USD) | Days since last order | Typical reorder interval, days |
|---|---|---|---|---|
| Alder | 240,000 | 216,000 | 14 | 7 |
| Borough | 180,000 | 108,000 | 45 | 60 |
| Crest | 120,000 | 66,000 | 21 | 21 |
| Delta | 60,000 | 42,000 | 10 | 30 |
| Total | 600,000 | 432,000 |
The intervals are supplied account history summaries, checked against complete order histories. Consolidate split shipments under their original order before estimating them. Where history is sparse or buying is irregular, show the interval as unavailable rather than pretending a reliable baseline exists.
Calculate contribution as revenue less variable costs. This follows the contribution margin definition in OpenStax's managerial accounting text; contribution remains available to cover fixed costs before profit.
Contribution = Net revenue − Variable costs
Cadence ratio = Days since last order / Typical reorder interval
| Account | Contribution (USD) | Contribution rate | Cadence ratio | Revenue rank | Contribution rank |
|---|---|---|---|---|---|
| Alder | 24,000 | 10.0% | 2.00 | 1 | 3 |
| Borough | 72,000 | 40.0% | 0.75 | 2 | 1 |
| Crest | 54,000 | 45.0% | 1.00 | 3 | 2 |
| Delta | 18,000 | 30.0% | 0.33 | 4 | 4 |
| Total | 168,000 | 28.0% |
Rates are rounded to one decimal and cadence ratios to two. Alder's contribution is 240,000 − 216,000 = $24,000; its rate is 24,000 / 240,000 = 10.0%. Borough contributes 180,000 − 108,000 = $72,000, so it outranks Alder on contribution despite smaller sales.
Alder's elapsed gap is 14 / 7 = 2.00 expected intervals. Borough's is 45 / 60 = 0.75. Calling Borough first merely because its last order was longer ago would ignore how these accounts buy. Crest is at its expected interval, and Delta is earlier in its cycle.
The contribution check is exact: 24,000 + 72,000 + 54,000 + 18,000 = $168,000, also 600,000 − 432,000. The total rate is 168,000 / 600,000 = 28.0%; do not average the account rates.
A cohort groups customers acquired in the same period. Wharton research on cross-cohort customer analysis defines acquisition cohorts this way and examines why behavior can differ between them.
For this distributor, define acquisition as the first completed order. Define repeat purchasing as another completed order within the following 90 days, excluding the initial order. A separate customer history extract gives these counts; all customers have completed their observation window by September 30, 2026.
| Acquisition cohort | Eligible new customers | Customers with a repeat order | Repeat purchase rate |
|---|---|---|---|
| Fourth quarter 2025 | 20 | 14 | 70.0% |
| First quarter 2026 | 25 | 20 | 80.0% |
| Second quarter 2026 | 10 | 6 | 60.0% |
| Total | 55 | 40 | 72.7% |
The first cohort is 14 / 20 = 70.0%. The combined rate is (14 + 20 + 6) / (20 + 25 + 10) = 72.7%, rounded to one decimal. Count customers once regardless of how many repeat orders each placed.
The latest completed cohort has the lowest observed rate. That is a reason to inspect acquisition channel, account fit and onboarding, not proof that one caused the difference. These are small groups, and seasonal purchasing can change the result. Compare contribution over the same customer age before deciding which cohort is commercially better.
Give Alder's account owner the cadence exception and the contribution detail together. The discussion should establish whether a reorder is due and why contribution is low. Borough needs a different conversation: protect a valuable relationship without assuming it is overdue.
For expansion, catalog fit per customer assesses which lines an account could plausibly buy; missing purchases alone do not establish demand. The adjacent RFM analysis for products applies purchase patterns to products and should remain separate from account priorities.
Record the chosen action, owner and next review date outside the numerical ranking. On the next review, check whether flagged accounts ordered and whether the action improved contribution. A useful measure must lead to a decision that can later be assessed.
Bring a sample of your own customer, order and cost exports to Covirage. Covirage maps the columns, its tools compute and check the account and cohort figures, and the external AI model explains the results with every figure traced to its rows. Create a free Covirage account; account creation takes about a minute, then you set up your business and bring your file.
See sales intelligence for the supporting business use.
Consider another method when the decision depends on costs, contractual renewal dates, seasonal buying or customer age. RFM summarizes purchase history, but those questions require additional evidence. Keep it as a descriptive view if useful, while using the relevant measure to set priorities.
No. Contribution measures revenue less defined variable costs over an observed period. Customer lifetime value estimates future economic value and depends on assumptions about future purchasing, retention, costs and timing. An observed contribution ranking does not establish a lifetime value ranking.
No. A longer gap is a review signal in a business without explicit cancellations. The customer may have inventory, a seasonal shutdown or a delayed project. Confirm the account situation before treating the absence of an order as a lost relationship.
Compare new customers within acquisition cohorts over equal observation windows. Keep them separate from established accounts when there is too little history to estimate a typical reorder interval. A new account can have low cumulative revenue simply because it has had less time to buy.