You sent the survey, the field period closed, and the number that comes back looks small. Before you decide whether that is a problem, you need to know exactly what the number measures. This article covers the basics of survey response rate: the formula, the related rates people confuse with it, and why a higher figure is not always a better one. It is part of our guide to customer surveys that produce decisions.
Key takeaways
- Survey response rate is the share of eligible people you invited who completed the survey, calculated as completed surveys divided by eligible invitations times 100.
- In the worked example, 920 completes from 4,600 eligible invitations is a 20.0% response rate, while dividing by all 5,000 sends gives 18.4%.
- Response, completion and cooperation rates answer different questions: completion points to the questionnaire, cooperation to the invitation and topic, and response rate mixes both.
- There is no single good response rate; compare against your own earlier waves with the same audience, mode and survey length.
- A high response rate is not proof of good data, because nonresponse bias can make a 22% rate overstate satisfaction by more than 20 points.
What is a survey response rate?
Survey response rate is the share of the people you intended to reach who completed the survey. The key word is intended. The denominator is the eligible people you invited, not everyone who happened to see a link.
The common shorthand, “the percentage of people who complete a survey,” leaves out the denominator. That is where most disagreements start: two teams can both report 20% and mean different things.
How do you calculate survey response rate?
The formula is:
Response rate = completed surveys ÷ eligible invitations × 100Eligible invitations means everyone you invited, minus people who could never have responded: email addresses that bounced, invitees who turned out not to qualify, and duplicates.
Worked example
These numbers are illustrative. A brand team emails a customer survey to 5,000 addresses.
| Step | Count |
|---|---|
| Invitations sent | 5,000 |
| Undeliverable (bounced) | 400 |
| Eligible invitations | 5,000 − 400 = 4,600 |
| Completed surveys | 920 |
| Partials (started, not finished) | 230 |
Response rate = 920 ÷ 4,600 × 100 = 20.0%.
If you had divided by all 5,000 sends instead, you would report 18.4%. Neither is wrong, but you must pick one definition and keep it the same from wave to wave.
What is the difference between response, completion and cooperation rates?
These three rates are often mixed up, and each answers a different question.
| Rate | Question it answers | Formula | Using the example |
|---|---|---|---|
| Response rate | How much of the eligible sample did we hear from? | Completes ÷ eligible invitations | 920 ÷ 4,600 = 20.0% |
| Completion rate | Of those who started, how many finished? | Completes ÷ starts | 920 ÷ (920 + 230) = 80.0% |
| Cooperation rate | Of those we actually reached, how many took part? | Completes ÷ eligible people reached | 920 ÷ 3,000 = 30.7% (if 3,000 opened the invitation) |
Completion rate tells you about the questionnaire: a low figure points to length, confusing questions or a poor mobile layout. Cooperation rate tells you about the invitation and the topic, because it removes people who never saw your message. Response rate gives the overall picture but mixes both effects.
Cooperation rate definitions vary between organizations, so say how you count “reached” when you report it.
What is a good survey response rate?
There is no single answer, and be wary of any article that gives you one number. What counts as good depends on the mode and the audience.
- Mode. Email invitations, in-product prompts, phone calls, intercepts and mail each draw different rates. A prompt shown to someone mid-task behaves very differently from a cold email.
- Audience. Customers who know you, employees, business buyers and a general-population list respond at very different levels.
- Relationship, topic, incentive and length. People respond more readily to a brand they use and a topic that affects them.
The most useful benchmark is your own: the same audience, mode and survey length in previous waves. A fall from one wave to the next tells you more than comparison to a published average.
Why is a high response rate not the same as good data?
A response rate describes how many people answered. It does not describe who answered. The risk that matters is nonresponse bias: the people who respond differ from the people who do not, on exactly the thing you are measuring.
Consider an illustrative customer base of 10,000, of whom 6,000 (60%) are satisfied and 4,000 (40%) are not.
| Scenario | Responding | Respondents | Satisfied share among respondents |
|---|---|---|---|
| A: satisfied customers respond at 30%, unsatisfied at 10% | 1,800 + 400 | 2,200 (22% response rate) | 1,800 ÷ 2,200 = 81.8% |
| B: both groups respond at 15% | 900 + 600 | 1,500 (15% response rate) | 900 ÷ 1,500 = 60.0% |
The lower response rate in scenario B gives the more accurate answer, because non-responders look like responders. Scenario A has the higher rate and overstates satisfaction by more than 20 points.
So treat response rate as a warning light, not a quality score. Compare your respondents with what you know about the population (region, tenure, plan, segment) and use weighting where it is justified. Our piece on why people lie on customer surveys covers a related problem, bias among those who do respond.
How can you improve your survey response rate?
Each of these works by lowering the effort or raising the reason to respond.
- Keep it short and focused. Every extra question costs completions. Cut anything you will not act on, and tell people up front roughly how long it takes.
- Make the invitation clear and personal. Address people by name, say who you are, why you are asking and what will happen with their answers. People respond more readily when they believe their view will be used. The psychology of surveying is worth reading here.
- Consider an incentive, carefully. A reported e-Rewards Market Research study found a $2 cash reward for completing a survey more effective than a $2,500 sweepstakes entry. Treat that as general guidance: guaranteed small rewards tend to beat long-shot prizes. Offer the same incentive to everyone, or you risk skewing who responds.
- Pick your moment. Send when your audience is likely to be free. For a business audience that may mean midweek, for consumers an evening. Avoid known holidays. Test rather than assume.
- Use more than one channel. Email, SMS, in-product prompts and social can each reach people the others miss. Mixing modes can also improve the balance of who responds. Keep subject lines free of spam-trigger words such as “free” or “free prize.”
- Send reminders to non-responders. Client Heartbeat has reported that two reminders can double a response rate. That figure is general guidance, not a promise; your result will depend on your audience. Send reminders only to people who have not yet responded.
If you collect open-ended answers, remember that they cost respondents more effort than a tick box. We cover the trade-offs in the benefits and challenges of open-ended survey questions.
Common mistakes with response rates
- Comparing rates across different definitions. One team divides by all sends, another by eligible invitations. The numbers are not comparable.
- Chasing the rate at the expense of balance. Heavy follow-up of easy responders can raise the rate while leaving the same groups missing.
- Treating a low rate as proof the data is useless. A low rate is a reason to check for bias, not to discard the results.
- Treating a high rate as proof the data is right. Scenario A above shows why not.
How do you do this in practice?
Once responses are in, you want to see who answered and whether the groups look right. Halo Reports from mTab supports crosstabs and banner tables with weighting and significance testing, so you can compare respondent profiles with your population and check results by segment. You can see the details on the Halo Reports page. From there, you can follow satisfaction and loyalty by journey.
Next time you field a survey, write down your definition of response rate before you launch, record sends, bounces and completes, and compare the respondent mix with your customer list when the data comes back. If you run a recurring program such as a customer satisfaction survey, those same records will let you track the rate wave by wave. mTab Pulse monitors trackers and studies and surfaces meaningful movement, so a shift between waves does not go unseen.