Customer lifetime value software: how to evaluate it

Customer lifetime value software is worth evaluating by whether your team can reproduce its customer contribution figures and understand its future assumptions. Start with a reconciled file, require the same customer population across comparisons, and choose the approach that fits the decision you need to make.

Define the decision before building a shortlist

An observed contribution report and a future lifetime estimate answer different questions. The first explains which customers have generated value so far. The second estimates what customers will generate after the measurement date.

Fader and Hardie's customer-base analysis tutorial defines customer lifetime value in terms of future cash flows and distinguishes it from historical customer profitability. A tool can calculate historical contribution correctly without establishing the assumptions needed for a future value estimate.

Write the decision in business terms: select a retention campaign, review costly accounts, set acquisition spending limits, or compare cohorts after the same follow-up period. Then write down the unit of analysis, observation window, contribution definition and owner. A general request to find your best customers leaves too many choices hidden.

The LTV calculation guide owns the formulas. This evaluation focuses on whether a product or working method lets your team apply and check those formulas consistently.

Compare approaches by the work you need to verify

There is no universal best category. Use the table as a shortlist for a practical test, rather than treating a feature list as evidence that customer value has been measured correctly.

Excel's SUMIFS function adds values meeting specified criteria, which makes it useful for a reference cohort calculation. Microsoft documents that Power BI visualizations can export summarized or underlying data, subject to permissions and report settings. Those documented capabilities are starting points for a test; neither establishes that your customer cost definition is correct.

Covirage is not affiliated with Microsoft, Excel or Power BI. Product names identify examples of the approaches above; selection should follow the observed test results for your files and team.

For recurring reporting requirements, use the separate customer lifetime value dashboard article. Keep that reporting design separate from the decision about how future customer value will be estimated.

Give every candidate the same reconciled file

Consider an illustrative US distributor's completed customer follow-up window. Its opening cohort contains four customers, including D, who places no follow-up orders. The table is the acceptance fixture; all costs use the same management contribution policy.

Net revenue is gross invoices less returns. Contribution is net revenue less product and variable service costs. In the displayed layout, enter these formulas in D2 and G2 and fill down through row 5:

=B2-C2
=D2-E2-F2

The proof is $2,450 - $150 = $2,300 net revenue and $2,300 - $1,220 - $260 = $820 contribution. Customer contributions also sum to $480 + $280 + $60 + $0 = $820.

Observed contribution per original customer is $820 / 4 = $205.00. Dividing by the three purchasing customers produces $273.33 instead. That second figure is valid only when labeled contribution per purchasing customer; it must not replace the original-cohort measure. Amounts per customer are rounded to two decimals after calculation.

The reference formula for contribution per original customer is:

=SUM(G2:G5)/COUNTA(A2:A5)

Require every candidate to reproduce the four customer results and the totals before evaluating additional capabilities. If a candidate imports transactions but omits the opening customer roster, explicitly supply that roster and repeat the test.

Test cost boundaries and future assumptions separately

Add acquisition cost only after agreeing on the displayed measure. The fixture's acquisition expense schedule provides the additional input below:

The $320 acquisition expense is charged once. The $500 result measures historical contribution after that spending, while $125.00 expresses the same amount per original customer.

Those are still historical figures. Renaming $125.00 as lifetime value does not make future retention or future orders known. The ledger-based lifetime value discussion explains what the transaction history can and cannot establish.

For a future-value function, ask the supplier to identify its starting date, customer population, forecast horizon, purchase assumptions, cost assumptions and discount timing. Ask what happens when the horizon ends. A value that includes an assumed tail beyond the displayed period needs that tail shown explicitly.

Use one-input sensitivity tests. Change retention while holding order frequency and contribution constant. Then change contribution while holding behavior constant. The result should move in the expected direction and the supplier should explain why. Do not accept a single precise estimate as a substitute for a range of plausible outcomes.

Use an acceptance checklist instead of a feature score

Record the evidence beside each requirement. A passed arithmetic fixture is necessary, but finance also needs a repeatable process when next month's file arrives.

Include the business owner in the test. Ask them to explain why customer C contributes less than customer A and whether D remains in the denominator. Their explanation should refer to the visible rows, not require confidence in an unexplained score.

After selection, use the customer value improvement guide to turn a measured driver into a costed business action. Buying software and improving customer economics are separate decisions.

Where evaluations go wrong

• Comparing different definitions. Revenue-based estimates and contribution-based estimates cannot be ranked as though they measure the same outcome.
• Accepting attractive totals without rows. A company total can match while individual customers carry duplicated revenue or missing costs.
• Losing inactive customers. Removing D changes the denominator and inflates the apparent cohort result.
• Changing several assumptions together. You cannot explain a sensitivity result if retention, costs and horizon all move at once.
• Checking only the initial file. Repeat the agreed reconciliation when the next period arrives and document any source or definition change.

Evaluate your customer file in Covirage

Upload a customer contribution file and opening roster; Covirage maps the columns, and its tools compute and check the figures. The external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data; creating the account takes about a minute.

See customer profitability for the business questions that support this evaluation.