Start with the decision, not the questions
Before writing the first question, write down the decision the results will feed. For example: keep or drop a menu item, offer delivery to more districts of Tashkent, find out why Instagram buyers do not come back, or choose between two packaging options.
Then test every question: what will you do differently depending on the answer? If nothing changes, cut it. Questions asked out of curiosity make the survey longer, and fewer people reach the end.
Who to ask
The sample depends on the goal. For service quality, ask recent buyers while the experience is fresh. For churn, ask people who stopped buying: they are harder to reach, but their answers tend to be more candid. For a new product, ask people from your target audience who have not bought from you yet.
Do not rely only on your Telegram channel subscribers or friends. They already like you, so the picture will be too warm. If customers come from different places, such as a store, Uzum, OLX or Instagram, do not pool them together. Compare the groups against each other.
How to write survey questions without leading the answer
The most common mistake is a question with the desired answer built in. “How much did you enjoy our fast delivery?” already calls the delivery fast. A neutral version: “How long passed between ordering and receiving, and how would you rate that from 1 to 10?”
The second trap is hypothetical questions. “Would you buy our new product?” almost always gets a yes, because the respondent risks nothing. Ask about real behavior instead: “How do you solve this problem today? Where did you last buy, and how much did you spend?” A past action tells you more than any promise.
Also avoid bundling. “Are you happy with the speed and quality?” is two questions, and an answer of “partly” explains nothing.
Rules for a survey people will finish
These look like small details, but they separate a useful survey from a useless one.
- One question, one topic.
- Plain words, without jargon or trendy terms the customer may not know.
- Answer options that do not overlap, including price ranges.
- Always include “Other” and “Hard to say”, otherwise people pick a random option.
- The same rating scale throughout the survey.
- One or two open questions at most: they are valuable but take effort.
- If your audience is mixed, prepare Russian and Uzbek versions.
Length and channels
The shorter the survey, the more people finish it. For most goals, 5–10 questions that take a few minutes are enough. If you have more topics, split them into two rounds or give different topics to different customer groups.
Pick the channel your customers already use to talk to you: a Telegram bot or chat message, a link in a newsletter, an Instagram story poll, a QR code at the checkout, an SMS link after purchase, or a phone call. Outside the capital and with older customers, a short call often beats a link, because people find it easier to talk than to fill in a form.
Motivation and honest answers
Open by explaining why you are asking and what will change as a result. A small bonus, such as a discount on the next order, raises the number of responses, but it also attracts people who answer only for the gift. Do not tie the reward to the rating and do not hint at the answer you want.
Anonymity helps too. When people do not fear offending a specific salesperson, they speak more directly.
What to do with the results
Count how answers are distributed, compare customer groups and read every open answer in full. Repeated phrases often point to a problem the numbers hide. Then go back to the decision you started with and write down specific changes.
Tell customers what you changed because of the survey. They will answer more willingly next time, because they can see their words were not wasted.
If the survey is part of a wider strategy, such as choosing positioning, reviewing products or finding growth points, the markdev team can help define the goal, build the questionnaire and turn the answers into an action plan. Describe your task and we will suggest where to begin.
FAQ
How many questions should a customer survey have?
For most goals, 5–10 questions that take a few minutes are enough. The shorter the survey, the more people finish it. If you have many topics, run two short surveys instead of one long one.
Which channel works best for a survey in Uzbekistan?
The one your customers already use: a Telegram bot or chat, a newsletter link, Instagram stories, a QR code at the checkout, an SMS link after purchase or a call. Phone calls work well with older customers and in the regions.
Should I offer a bonus for completing the survey?
A small bonus increases responses but can attract people who answer only for the reward. Do not link the reward to what people say, and explain how the results will be used.
How can I tell that a question is leading?
If the wording contains an opinion (“excellent service”, “fast delivery”) or hints at the right answer, it is leading. Replace it with a neutral question about facts: what the customer did, when and how much they spent.
Let’s look at your task for free
Leave your contact — in a 15-minute call we’ll show where leads get lost and what to fix first.