Customer Feedback Tools: What to Check Before You Buy

Take the category label away and three jobs are left. Every product does some mix of them, and the mix is what you are choosing between.

• Ask. Getting a question in front of a customer at a moment when answering it is easy.
• Collect. Storing the answers somewhere they can be counted, rather than read once and lost.
• Act. Routing what came back to the person who can do something about it.

Most demos spend their time on the middle job, because dashboards photograph well. The first is where tools actually differ, and the third is where they quietly fail: a platform that cannot pass an angry answer to a human has built a very expensive complaints archive. Keaz sits on the ask-and-act side of this — it sends the request and handles the reply on WhatsApp for Shopify stores, and what that covers is deliberately a short list. It is not a survey platform, and it does not host reviews.

Feature lists are written to survive comparison. These four questions are not.

1. What is the response rate on the channel it uses? Ask for the number, not the industry claim. The channel decides most of it — a question inside a conversation the customer already has open behaves differently from one in an inbox they check on Sundays.
2. Which events can trigger the ask? A tool that can only send on a schedule can only ask generic questions. One that fires on a delivery, a second order or a support reply can ask about the thing that just happened.
3. What consent does it need, and do you have it? This is the reachable-share number from the introduction. It is the one most likely to be quietly awful, and the one a demo will not raise on your behalf.
4. Where does an unhappy answer land? Some share of feedback is a complaint wearing a rating. If it lands in an analytics table instead of in front of a person, you have automated your way out of the recovery.

If a vendor answers three of the four and talks around one, the one they talked around is the answer. What the channel looks like by vertical is a better guide to fit than any feature matrix.

After the thing you are asking about has finished happening, and not a week later. That sounds obvious, and it is the most common configuration error in feedback programmes: the request fires on the order event, arrives while the parcel is still in transit, and asks a customer to rate a product they have not seen.

The mechanics in Keaz are deliberately narrow. A flow starts from the Contact purchase Shopify event, a Delay block holds the message until after delivery, and a Messages block carries your existing review or feedback link as a button. The Contact Purchase trigger documents which filters it offers — product name, product category, order count — so the ask can be specific rather than generic. For customers who already bought, there is a second route: build a segment of the people who bought one product and send them a single newsletter with the same button. Both routes are written up in collecting reviews with Keaz.

Keaz does not store the answers. You link out to wherever your reviews or responses already live, and the channel delivers the ask. If you are collecting on Google specifically, the step-by-step version covers the link format and the rules Google applies to review requests.

More than the send fee. Every feedback request spends a small amount of the permission you hold, and permission does not refill on its own. Baymard Institute's 2026 quantitative study of 1,083 US online shoppers found that 48% of them unsubscribe from a retailer because the emails arrive too often from that specific retailer — ahead of general inbox clutter at 46% (Baymard Institute, 2026, retrieved 2026-09-24). Frequency, not content, is now the leading reason a list shrinks. A quarterly survey layered on top of your existing campaign calendar is a frequency decision before it is a research decision.

The second cost is consent. Under the GDPR, consent is a freely given, specific, informed and unambiguous indication of the data subject's wishes (Regulation (EU) 2016/679, Art. 4(11)) — so permission collected for order updates is not automatically permission for a research programme, and a tool that imports a customer list without raising that question is handing you a problem rather than a feature. In Keaz the flow filters on the contact's opt-in level before it sends, which answers the question at send time rather than in an audit. This is a summary rather than legal advice — have your own counsel review your setup.

• Asking before the product arrived. The request fires on the order rather than the delivery, and you collect ratings for a parcel in transit. Add a delay and set it to your real delivery time, not your promised one.
• Asking twice. A repeated request reads as pressure and costs opt-ins. One ask per event, and no second attempt on the people who did not answer.
• Buying for the dashboard. Charts are the cheapest part of this category to build. Response rate and routing are the expensive parts, and they are the parts a demo skips.
• Treating a complaint as a data point. An unhappy answer is a recovery opportunity with an expiry date. If nobody sees it inside a day, the tool has converted a saveable customer into a statistic.
• Measuring satisfaction and nothing else. A rising score next to a falling repeat-purchase rate means the people who stayed are happy and the rest left quietly. Read it alongside the two retention numbers that matter.