Why Customers Buy: Find Your Real Reason in Your Data

Why do customers buy from your brand?

Customers buy a product to make progress in a particular situation, and the product is only the tool they pick for the job. The Christensen Institute describes Jobs to Be Done as a lens on the circumstances that push people toward and away from decisions, and says people "hire" products and services when a job arises in their lives. It sets this against marketing that focuses on demographics and product attributes.

For an ecommerce brand, the useful part is that the job is rarely written on your product page. It is written in what customers say after the sale, in what they ask before it, in what they send back and in when they return. Those records already sit in your helpdesk, your returns data and your order history.

Where is the real reason customers buy recorded?

The reason customers buy is recorded in four places you already own: reviews, support tickets, return reasons and repeat-purchase behavior. Each one catches a different moment in the buying cycle, so use all four.

• Reviews: customers describe the outcome in their own words, right after use. Look for the situation they mention, such as a trip, a gift, a new baby or a repair, more than the star rating.
• Support tickets: before-purchase questions show what customers are unsure about, and after-purchase questions show what they expected. Gorgias lets you export ticket data to a CSV file from any view, and with message content for the most recent 30 days from its analytics reports.
• Return reasons: a return is a job the product failed to do. Shopify's self-serve returns lets you customize the request form to collect return reasons and product condition, so add the question if it is missing.
• Repeat behavior: when and what customers reorder shows whether the job comes back. Shopify's Returning customers report lists customers with two or more orders, and the Customer cohort analysis report groups customers by first order date and shows how often each group returns.

How do you read tickets and reviews as jobs?

You read a ticket or review as a job by rewriting it as one sentence: the situation the customer was in, what was getting in the way, and the result they wanted. Then you count how many customers share each sentence. Work through it in this order:

1. Pull a sample of 100 or more recent reviews and tickets, mixed across your top products. Export tickets through your helpdesk's export and reviews through your review app.
2. Remove noise first. Automatic order confirmations that land in the helpdesk and spam replies are not customer language. Why order confirmation emails show up as helpdesk tickets explains how that inflates counts.
3. Write a job line for each real item: "When [situation], I wanted [result], but [obstacle]." Keep the customer's own words inside the brackets.
4. Group lines that describe the same situation, and count customers, not tickets. A customer who wrote six times counts once.
5. Rank the groups. The top two or three are your candidate jobs. Note which ones come from your best customers, meaning the ones who ordered again, because their job is the one worth protecting.
6. If an AI model does the tagging, label the result model-read and check a random 20 items by hand before you trust the counts.

What do returns and reorder patterns add?

Returns and reorder patterns show the job from the outside: what went wrong, and how often the need comes back. Reviews can flatter a brand, while a return reason and a reorder date record what customers did.

Read them together. A product that is returned for the same reason again and again was bought for a job it does not do, and the fix is usually in the page, the sizing help or the expectation you set. A product that customers reorder at a steady interval is doing a recurring job, and the interval tells you when a reminder is welcome. An order that never repeats may belong to a one-time job, such as a gift or an event, and a retention discount will not change that.

Treat these as patterns to test, not proof. A cohort that returns more often could reflect the season, a promotion or the product mix. Check the same cut across two different months before you act on it.

How do you check that your messages match the job?

You check the match by putting your messages next to the customers' words and looking for the gap. Open the homepage headline, the best-selling product page, your best ad, the order confirmation, one email flow and one text flow, and mark every phrase that describes a feature or an ingredient instead of the job.

• Reuse the phrases customers repeat. If many reviews say a product made a task faster, then "faster" is your headline, not the spec behind it.
• Answer the top before-purchase ticket on the page itself. A question customers keep asking support is a gap in the page.
• Match retention messages to the job. A reminder tied to the usual reorder gap fits a recurring job. A discount code fits less well for a customer who bought for the story or the look.
• Keep one lead message, and let supporting messages vary by customer. People can buy the same product for different jobs, so emails and texts may stress different reasons without changing the brand's main promise.

How do operations deliver on the job?

Operations deliver on the job through the parts of the experience a customer feels after clicking buy: dispatch speed, packaging, tracking, support answers and the return process. Each job puts a different load on those parts, which is why positioning and operations are one problem.

• Job: have it by a date. Then the delivery estimate, the cutoff time and proactive tracking updates are the product. A late parcel with no warning breaks the promise even when the item is perfect.
• Job: get it right without a trial. Then sizing help, clear return rules and a quick exchange matter more than another feature line. Check how long a return takes to process.
• Job: never run out. Then reorder reminders, subscription handling and stock availability are the product. A recurring customer who meets a stockout has been failed on the one thing they came for.
• Job: a gift or a treat. Then packaging, the unboxing and a clean order status page carry the message, and a support reply must read as part of the same brand.

Sources

Every outside fact in this guide was read on the source's own page on November 11, 2026. Reports and menus change, so check the page before you rely on a step.

• Christensen Institute: Jobs to Be Done. Source for the definition of the lens and the idea of hiring products.
• Gorgias Docs: Export tickets from Gorgias. Source for the CSV export from a view and the export with message content.
• Shopify Help Center: Self-serve returns. Source for collecting return reasons on the request form.
• Shopify Help Center: Customers reports. Source for the Returning customers report and the Customer cohort analysis report.

Frequently asked questions

How many customers do you need to find the real reason they buy?

There is no fixed number, but a sample of 100 or more recent reviews and tickets, with each customer counted once, usually shows whether two or three reasons dominate. If the groups stay scattered after that, add more data from a different month before drawing conclusions. A pattern that appears in only one month may be seasonal.

Can the reason customers buy differ by product?

Yes. One brand can sell a repair product for an urgent job and a gift set for an occasion, and the same product can be bought for different jobs by different people. Run the reading separately for your top products, then look for a shared thread at the brand level.

How often should you repeat this review?

Repeat it when the evidence changes: after a new product line, a change in return rate, a spike in one ticket reason or a shift in who is buying. A light check each quarter is enough for most stores. Between reviews, a tagged feed of ticket reasons shows early drift. Morsegrid's Insights product is built to sort conversations into themes.