Net Promoter Score for Online Stores

Net Promoter Score (NPS) measures how likely customers are to recommend your online store, using a question scored from 0 to 10. You calculate it by subtracting the percentage of detractors from the percentage of promoters. For a store, a useful score comes with the reasons customers gave it and a consistent survey method.

Ask after customers receive their order so their answer includes delivery and the product itself. A rising score is more useful when the comments show which problems you fixed.

What Net Promoter Score Measures

The NPS question measures a customer’s stated likelihood of recommending your store to someone else. The survey asks, “How likely are you to recommend our store to a friend or colleague?” Customers select 0 for not at all likely through to 10 for extremely likely.

Bain’s scoring method divides responses into three groups:

• Promoters give 9 or 10.
• Passives give 7 or 8.
• Detractors give 0 to 6.

These are score categories. A detractor hasn’t necessarily complained publicly, and a promoter hasn’t necessarily referred a friend. Their comments and later purchases help you understand what the rating means.

Fred Reichheld at Bain & Company introduced NPS in 2003. For an online store, the recommendation question can reflect the whole purchase, including whether the product matched its description.

Customer satisfaction (CSAT) usually asks how satisfied someone was with a particular experience, such as a support conversation. NPS asks about recommending the store. A helpful refund conversation can receive a good CSAT rating while the customer still gives a low NPS after receiving the wrong item.

How to Calculate NPS

Calculate NPS by subtracting the percentage of detractors from the percentage of promoters, using all valid scored responses as the total.

NPS = percentage of promoters - percentage of detractors.

In this hypothetical online-store survey, 200 customers respond:

Response group	Rating	Responses	Share of all responses
Promoters	9 or 10	120	60%
Passives	7 or 8	50	25%
Detractors	0 to 6	30	15%
Total	0 to 10	200	100%

The calculation is 60 minus 15, giving an NPS of 45. Report it as 45, with 200 responses, rather than as 45% satisfaction.

Passives stay in the denominator even though you don’t subtract their percentage. Excluding the 50 passives would change the total to 150 and incorrectly inflate the result to 60.

The score runs from minus 100 when everyone is a detractor to 100 when everyone is a promoter. An NPS of zero means the promoter and detractor percentages are equal, even if most responses are passive.

For Excel, put valid numeric ratings in cells A2 to A201. This formula returns the score for those 200 rows:

=(COUNTIF(A2:A201,">=9")-COUNTIF(A2:A201,"<=6"))/COUNT(A2:A201)*100

Validate the column first: every filled cell must contain a number from 0 to 10. Keep blank responses out of the calculation and check the result against your group totals.

Running an NPS Survey After an Online Order

Send an online-order NPS survey after confirmed delivery, with enough time for the customer to inspect or use the product. At checkout, the customer can judge ordering but hasn’t experienced delivery or what arrived.

For ordinary retail orders, start with a survey three to seven days after delivery. Adjust that delay for products that need longer use, then keep it consistent within the category. Treat this as a starting schedule for your store, not an NPS requirement.

Use the same recommendation question and scale every time. Follow it with, “What is the main reason you chose that score?” Make the written answer optional so customers can submit their rating without writing a comment.

Set a repeat-contact rule before switching on automated sends. A practical starting rule is one NPS invitation per customer every 90 days, even when they place several orders. You can still offer separate feedback after a support conversation. Keep those interaction ratings separate from the NPS series.

Include customers with poor experiences in the same eligible order group. Don’t filter out a shopper because they contacted support or requested a return. Record invitations sent and responses received so a score doesn’t conceal a falling response rate.

There’s no response count that automatically makes NPS reliable. For a small store, use 100 responses as a practical first reporting milestone and treat smaller samples as provisional. This is a reporting choice, not a statistical guarantee. With just 20 responses, moving one rating from detractor to promoter changes NPS by 10 points.

If you receive few responses, combine consecutive periods with the same survey method and show the total response count. Read comments immediately, even before reaching your reporting milestone. Choose NPS survey software that can send after delivery and enforce your contact limit. You also need to export scores with their comments.

What is a Good Net Promoter Score

A good NPS depends on your industry and how you ask the question. A score above zero means promoters outnumber detractors among respondents. It doesn’t establish that your store outperforms similar retailers.

A checkout pop-up and an email after delivery reach customers at different moments. A general brand survey also asks something different from a survey tied to an order. Compare published benchmarks only when their customer group and collection method fit yours.

Bain’s guide to reliable NPS measurement explains how survey methods and response bias can change the result. Your store’s own trend is more useful when the question, timing and eligible customers stay consistent. Show response counts beside each score, and mark method changes in your reporting.

NPS has limits you can see in the calculation. It groups a rating of 0 with a rating of 6, so improvement within that group leaves the score unchanged. A move from 8 to 9 changes someone’s category after only a one-point rating increase. The single number hides those differences.

It also measures stated intent. Customers can recommend your store while buying elsewhere, and people who answer can differ from those who ignore the survey. A low score alone doesn’t tell you which part of the order failed.

Pair NPS with CSAT after support interactions to see both the store relationship and the quality of help. A low NPS beside high support satisfaction suggests the team handled a problem well, but the underlying purchase still disappointed the customer. The guide to support metrics explains how to report those measures together.

Reading Detractor Comments Beside Chat Transcripts

Match detractor comments with relevant chat transcripts to find the part of the shopping experience that needs repair. Use an order reference or customer-approved identifier where available. An anonymous response can still show a theme, but you can’t reliably assign it to one conversation.

shoppa is an AI shopping agent that helps customers find, compare and buy products in the chat using your catalogue and documents. Its workspace includes a step-by-step trace of each conversation. Read that trace beside the customer’s survey comment to see what they asked and which answer they received.

In a hypothetical sizing complaint, the customer writes that the garment was smaller than expected. The transcript shows they asked for measurements, but the answer repeated a general size label. Check whether the product page and the agent’s source document contain the measurements the customer needed. Add the missing information where it belongs.

For a delivery complaint, compare the answer in the chat with the delivery policy and the actual order. If the policy omits regional delivery conditions, correct the policy and the agent’s source document. If the published estimate was accurate and the parcel arrived late, give the delivery issue to the fulfilment team.

Group repeated causes across comments instead of treating each low rating as the same problem. Assign a person to each fix and record what changed. The voice of customer guide explains how to turn those themes into ongoing work.

After a fix, try the affected question in your private test agent and confirm it answers from the corrected information. Watch the next comparable NPS survey for that complaint, alongside the score and response count. Use the result to choose the next fix, without assuming one score change proves what caused it.