Customer analytics for new product development

Customer analytics for new product development tests commercial assumptions against purchases your company has actually recorded. Compare related purchases by segment, offer the proposed product to a defined group, and measure purchase adoption over a consistent period.

The result should guide the next product decision: which customer group to interview, where to expand a pilot, or which assumption needs another test. Historical spending alone cannot prove customers will buy an offer that did not exist during that history.

Turn the product idea into a testable question

Suppose a US packaging supplier is considering a returnable shipping tote. Its existing core products are shipping supplies; packing inserts are an adjacent category. The team believes the segment with broader insert purchasing will show stronger tote adoption.

Write that assumption down before looking at results. Define eligible customers, the offer, the purchase period, and the event that counts as adoption. Here, adoption means a recorded pilot purchase with positive net pilot revenue, after returns and credits; it does not establish payment collection.

The Small Business Administration's market-research guidance distinguishes demand questions from broader market size, competition, and pricing research. Your own transactions can inform the demand test without answering those other questions.

Use customer segmentation from the ledger for the underlying group definitions. Keep those definitions fixed during the pilot so a customer moving between labels does not change the apparent result.

Prepare one customer row with explicit dates

Start with transaction lines containing account ID, product category, posting date, net revenue, and currency. Bring a customer list containing the segment and pilot eligibility. Keep an offer log showing that eligible accounts received the same offer and had equivalent time to respond.

Summarize purchases separately for the history and pilot periods. For this example, the history is April 1 through June 30, 2026, and the pilot is July 1 through September 30, 2026. All eligible customers received the offer on July 1. All amounts are USD, and no returns remain outstanding at the cutoff in this example.

The resulting customer worksheet uses columns A through E and rows 2 through 7:

A zero means the complete file contains no net purchase in that category and period. A missing export or unresolved account match should remain an exception rather than becoming zero. Defining the customer denominator covers why this distinction changes the comparison.

Calculate adjacent buying and pilot adoption

Use the same eligible customers in both segment denominators. Count accounts with positive adjacent revenue, then count accounts with positive net pilot revenue.

Adjacent-buyer rate = Eligible accounts with adjacent purchases / Eligible accounts

Purchase adoption rate = Eligible accounts with net pilot purchases / Eligible accounts

In Excel, put Distributor in G2 and Service company in G3. Use these formulas for eligible accounts, adjacent buyers, and pilot adopters:

=COUNTIFS($B$2:$B$7,G2)

=COUNTIFS($B$2:$B$7,G2,$D$2:$D$7,">0")

=COUNTIFS($B$2:$B$7,G2,$E$2:$E$7,">0")

Microsoft documents COUNTIFS for counts meeting multiple criteria. If those outputs sit in H2, I2, and J2, the adjacent rate is =I2/H2 and purchase adoption is =J2/H2; fill down for the other segment.

Percentages are rounded to one decimal place. The total divides three buyers by six eligible accounts; calculate it from the underlying counts. For segment pilot revenue, use:

=SUMIFS($E$2:$E$7,$B$2:$B$7,G2)

Microsoft's guide to worksheet sums explains using conditional sums for grouped amounts. Covirage is not affiliated with Microsoft.

Prove the total, then interpret the finding

The customer rows reconcile to segment revenue: $600 plus $1,500 equals $2,100. The adopter counts reconcile too: one distributor plus two service companies equals three purchasing accounts. Three adopters plus three nonadopters equals the six-account eligible cohort.

The observed result does not support the segment-level assumption in this pilot. Distributors had broader adjacent purchasing, but service companies had more pilot adopters in this group. Review the individual customers and interview nonbuyers before changing product plans. These are observed results for this cohort, not estimates of the broader market.

Test sensitivity to one additional purchase. If P02 records a net pilot purchase of $600, with every other input unchanged, distributor adoption becomes two divided by three, or 66.7%. Total adoption becomes four divided by six, or 66.7%, and net pilot revenue becomes $2,100 plus $600, or $2,700. That change shows how strongly a single customer can move a small pilot.

Assign ownership clearly. Product owns the hypothesis and decision, sales operations owns eligibility and offer evidence, and finance checks recorded revenue and credits. A new product development dashboard can organize subsequent reviews once the measures are defined.

Where it goes wrong

• Requests become purchases. Interest, an interview comment, an accepted quote, and a paid order are separate evidence. Keep each count labeled.
• Only contacted accounts appear. Record eligibility and outreach separately; a segment offered the product more often is not directly comparable without that context.
• History overlaps the test. Define the historical evidence before the pilot starts so adoption does not help determine the segment it is supposed to test.
• Revenue becomes contribution. The table shows net sales. Add validated product and service costs before drawing conclusions about profitability.
• A small cohort becomes a market forecast. The sensitivity example is a reason to extend the test, not multiply its adoption rate across an invented population.

Test your own product evidence in Covirage

Bring a sample of your own purchase export and customer file with segment and pilot eligibility; Covirage maps the columns, and its tools compute and check the comparisons. The external AI model explains the results with every figure traced to its rows. Create a free Covirage account, then set up your business and bring your file.

See sales insights for the supporting commercial workflow.