What belongs in the denominator of a Shopify post-purchase poll?

Your Shopify store booked 1,000 orders and received 120 answers to a thank-you page poll. The growth meeting starts with someone calling that a 12% completion rate. Before anybody rewrites the question, ask whether the 1,000 included retail orders, whether the poll ran all week and whether it was shown only to first-time buyers. Those details determine what the percentage can tell you.

Give each count a name a teammate could reproduce. Start with one survey, one placement, one time zone and one date range. Keep an order as the unit from start to finish. A customer placing two eligible orders creates two opportunities under that definition. Refreshing one order five times still creates one.

Work backward from the 120 answers

Consider a hypothetical store that ran an online, first-order survey all week. Of 1,000 total orders, 200 were retail orders outside this placement and 200 more were repeat online orders excluded by its targeting. The remaining 600 qualified for the question. The team has separate, reliable instrumentation showing that the survey rendered on 450 distinct eligible orders. It recorded a completed answer on 120 of those orders.

• Storewide yield is 120 / 1,000, or 12%. Use it to estimate how many answers the business gets per order across channels.
• Eligible-order yield is 120 / 600, or 20%. Use it to track the survey's contribution within the audience you intended to reach.
• Recorded display coverage is 450 / 600, or 75%. The 150-order gap is a reason to investigate the path and measurement. A recorded render also cannot tell you whether a buyer scrolled to the question.
• Completion per recorded display is 120 / 450, or about 26.7%. Use this definition only when the completed orders and recorded displays can be matched and counted consistently.

If 30 of the 120 respondents picked “Podcast,” then 25% of respondents chose that channel. You know 30 surveyed orders named a podcast. Ninety respondents named something else, and 480 eligible orders supplied no completed answer. The discovery sources of those 480 orders remain unknown.

Put your counting rules next to the chart

Write down when the survey became active, its order filters, the time zone, the eligible sales channel and how you handled test orders. Set a cutoff for late answers. For example, group orders by the day they were placed and allow 48 hours for answers to appear before closing each daily total. Apply the same window to every comparison. If you simply divide answers created today by orders placed today, a late response can land in the wrong group.

If the survey also appears on the Order status page, decide whether you can identify the placement for each submission. OrderSurvey’s Responses list shows the placement. In the Responses CSV, the Location column identifies where an answer was submitted. If your analysis source drops that field, keep thank-you and later order-status responses together in a clearly labeled combined total. Shopify explains when customers move from the Thank you page to Order status. The placement comparison covers the questions that fit each visit.

Diagnose a change before editing the poll

Suppose eligible-order yield falls from 20% to 15%. Check first whether targeting or the customer mix changed. If measured coverage fell and completion among displayed orders held steady, investigate why fewer eligible orders recorded a display. If coverage held steady, look at question length, changes to surrounding content and answer options. The survey wording experiment shows how to test one proposed change. With order counts and responses alone, both explanations remain possible.

Use OrderSurvey on the Shopify App Store to set up the question and review answers. Bring independently verified exposure data to this worksheet only if you have it. The survey benchmark guide is easier to interpret once everyone knows which rate appears on the chart.

Frequently asked questions