Feedback after a delivery tells you how one job went. It does not tell you how a customer feels about working with you after a year of orders, invoices, delays and good days. Without that wider view, small companies rely on impressions: the loudest complaint, the friendliest customer, the last conversation.
A yearly survey of the whole customer base fills this gap. It is short, it has a named owner and it ends in a few visible changes. It complements, and does not replace, the routines in collecting customer feedback in three questions and asking customers for feedback after delivery. Those catch problems as they happen; the yearly survey shows patterns.
Decide what the survey is for
Write one sentence before you write any question: "We run this survey to decide what to improve next year in how we sell, deliver and support." If a question does not help that decision, leave it out. Do not use the survey to sell, to collect data for another purpose or to prove that customers are happy. A survey built to flatter you produces flattering answers and no information.
Choose who receives it. For most companies this is every customer who bought in the last twelve months, plus important customers who did not. Include the ones you suspect are unhappy. Their answers are the most useful.
The five questions
Five questions take about three minutes. Longer surveys lose answers, and the people who finish them are not typical.
- Overall, how satisfied are you with working with us over the past year? Use a scale of 1 to 5. Keep the scale identical every year so you can compare.
- How likely are you to recommend us to a colleague or another company? Use a scale of 0 to 10. This is the familiar "net promoter" style question. Treat it as one signal among several, not as a score to chase.
- What is the one thing we do well that you would like us to keep doing? Free text. It protects what already works from being "improved" away.
- What is the one thing we should improve first? Free text. Asking for one thing forces a priority.
- May we contact you about your answers? Yes or no, with a contact field if yes. This lets you close the loop with the people who have something to say.
The GOV.UK Service Manual gives practical guidance on measuring user satisfaction, including why a consistent question and scale matter when you compare results over time.
Add at most one extra question that identifies the segment, for example the product or service the customer mainly uses. It lets you read the answers by group without asking people to repeat information you already have.
Who sends it
Send the survey from a person the customers know, such as the owner or the account manager, not from a no-reply address. Write two or three sentences: why you ask, how long it takes, and that you read every answer. Use your normal customer email address and a plain link to the form.
Check the rules that apply to you. Whether a survey invitation counts as marketing depends on its content and on your country; in the UK, the ICO publishes guidance on direct marketing and electronic communications. Keep the invitation free of offers, and if in doubt, ask whoever handles data protection for you.
Choose a quiet time. Avoid the weeks around year end, holidays and your customers' busiest season. Many companies pick one fixed month, for example the second month of the year, so it becomes a habit rather than a project.
One owner reads every answer
Name one person who owns the survey from invitation to action list. This person:
- sends the invitation and one reminder after about a week;
- reads every free-text answer, not only the averages;
- replies within a few days to anyone who asked to be contacted, especially low scorers;
- writes the summary and the action list.
Do not hand the raw answers to the whole team to interpret. A short summary from one reader is read; a spreadsheet of 80 rows is not.
Turn the answers into three actions
Spend one hour with the answers and sort them.
- Compare the numbers. Look at the average rating and the recommendation question, and compare them with last year's if you have it. Note any difference between customer groups.
- Group the free text into themes. Typical themes are delivery times, communication, price, invoicing and quality. Count how many customers mention each.
- Pick at most three actions. Choose the themes that are mentioned most or that matter most to your best customers. For each, write what changes, who is responsible and by when.
- Mark what you will not do. Some requests are not worth acting on. Say so in the summary, with a reason, so the decision is visible.
Three actions that get done beat ten that fade after a month. Add the actions to your normal planning, and check progress in your regular review, such as a one-page monthly operations review.
Tell customers what you did
A few weeks after the survey, send a short message to everyone you invited: "You told us X and Y. We are changing A and B by this date." Customers who answered see that it mattered. Those who did not may answer next year. It is also the cheapest way to show attention, because it costs one email.
Call the customers who gave low scores and agreed to be contacted. Listen first, thank them and say what you will do. Often a single conversation repairs more than any change in process.
Keep it simple the next year
File the questions, the summary and the action list in one folder per year. Next year, send the same five questions and compare. Resist the urge to add questions: the value of a yearly survey is the trend. If a new topic matters, ask about it in conversation or in feedback after delivery, and keep the yearly questions stable.
After a few rounds you will know which parts of your service customers value, where they struggle and whether your changes made a difference. That knowledge is difficult to get any other way, and it takes about three hours of work in a year.
