All Collections
Creating surveys
Setting up your audience
The sample size calculator: getting your sample size right
The sample size calculator: getting your sample size right
Updated over a week ago

Stuck with what sample size you need to use for your survey? Check out this handy guide for everything you need to know!

The most important three sample size questions to ask yourself:

Before we get going, keep in mind these three core principles when considering your sample size:

1. What are your subgroups of interest?: A lot of your business reporting might be on the overall sample you survey with Attest, but consider what you need when you drill down into results.

What are the target groups you’re interested in within your data? Male? Female? Brand users? Brand considerers? If you don’t know how many you will get, trying to estimate what percent of the population these could represent will help you make sure up front you are going for the right overall sample size.

2. Is your audience feasible?: Many brands will find their audiences are easily targetable with Attest (hooray!).

However, where we have a smaller panel audience in certain markets (this is often dictated by population size) or if you’re after more niche audiences (e.g B2B or people who own purple glasses), then you may not be able to use sample sizes that are as large as you would typically like.

You're likely to still be able to achieve a robust sample size in many instances, but contact the Customer Research Team if you are unsure about how feasible your qualification criteria are, and for guidance on the relevant sample size to use.

3. What is your use case?: Typically different use cases demand different levels of robustness. For example, brand tracking will have larger sample sizes than concept testing due to the nature of how the data is used (i.e. diving into subgroups of interest to inform business decisions).

If you need surveys with a regular frequency (e.g. brand tracking), multiple surveys launched at once (concept testing), or if you need results back by a certain date, contact the Customer Research Team to understand what’s feasible.

The art and science of sample size calculation

It’s a beautiful thing! The ‘art’ (as laid out in our core principles above) takes into account all the context of your target audience upfront. The ‘science’ is understanding what levels of robustness and accuracy you can expect from your sample.

Below we'll get into some of the nitty-gritty of sample sizes, but if you're looking for a more straightforward explanation, head over to our Consumer Research Academy.

Let’s start by exploring the scientific bit.

Using the calculator:

The calculator is your scientific best friend.

You can find it located on the Audience page, underneath the 'survey respondents' box.

Once you click on the link, the calculator modal will pop out.

  • Population Size: We’ve pre-programmed population sizes of popular Attest markets, so feel free to use these. If your market isn’t on the list, use ‘Custom’.
    Alternatively, if you have an idea of how many people are in your overall population then use ‘Custom’ from this drop down for a more accurate calculation.

  • Margin of Error: You can choose the level of accuracy you’d ideally want within your results. The percentages here indicate the maximum amount you’re happy for your results to differ from the actual population by. Lower margins of error require higher sample sizes.

  • Confidence level: We’ve preprogrammed standard confidence intervals used in market research. The higher the confidence level you choose, the more sure you can be that the data is representative of the population.

Once you're happy with your sample size, you can click `Apply to survey` and it will automatically update your sample size for you.

You can read more about the maths behind the calculator at the bottom of this page!

Thinking about your use case:

You’ve read about the science, now more about the art! This bit is often based on what the purpose of your research is. Here are more detailed sample size considerations for typical market research purposes:

1. Brand tracking

Helps you answer: how is my brand doing?

Sample size considerations:

  • Go Large: We would typically recommend n=1000 as brand trackers are typically based on bigger audiences to reduce the level of natural flux in data. This is why we recommend smaller margins of error in our calculator for brand tracking.

  • Break it down: What subgroups are you interested in e.g. will your business want to know what target females aged 18-34 think about your brand? Make sure you account for how these might fall out given your key questions.

  • Think long term: If you are running a brand tracker aim to keep your sample size consistent over time. Running brand trackers usually means applying some form of fresh sample each month/quarter and excluding anyone who has taken a recent previous survey which may impact the available audience size.

  • Is it feasible: The more niche your audience is, the more likely you are to need a smaller, but as robust as possible, sample size.

2. Concept testing/Creative testing

Helps you answer: which concept/creative should I launch and why?

Sample size considerations:

  • Monadic or Sequential monadic?: That is the question! Monadic testing, which is when you place each individual concept/creative in separate surveys (‘cells’), is on the whole a less biassed and deeper way of understanding the performance of each concept. Sequential monadic is when multiple concepts/creatives are placed in the same cell - you might do this if you have more niche audiences which mean many tests are not possible. This method is less effective when concepts/creatives are similar, or when you are testing many (e.g. more than two) concepts in a single test.
    We recommend where possible to carry out monadic testing so you can get the most unbiased data on your concept/creative

  • Quick tests: Traditional research agencies might only recommend n=100-150 on sample sizes. This is due to them taking much longer to fill surveys. At Attest, we recommend n=250 per cell because we can fill surveys much quicker which gives you robustness and confidence in the results you get back! How great is that?

  • Is it feasible? You’re likely to need results back quickly, so take this into account particularly where you are looking at more niche audiences. As with all other studies, consider how hard it is to reach the target audience and take guidance from our Customer Research Team if needed.

3. Price testing

Helps you answer: what’s my ideal price point?

Sample size considerations

  • Analysis needed: The type of analysis you want to run will dictate the type of study you want to run.

  • Is it feasible: Particularly where you are targeting a more niche audience, get in touch with the Customer Research Team with any queries.

4. Consumer profiling

Helps you answer: what are the attitudes and behaviours of my (potential) audience?

Sample size considerations

  • Go large: You’ll want a larger sample size because a lot of the time you might not know what subgroups of interest (‘personas’) might fall out from this type of study. The aim is to have overall robustness so you can dig into the data and understand how certain groups behave or think.

  • Is it feasible: Particularly where you are targeting a more niche audience, get in touch with the Customer Research Team for guidance on ideal sample sizes for your study.

We know there are a few considerations for choosing the right sample for your study. While this quick guide gives you key points to take into account, the Customer Research Team is always here to help! Don’t hesitate to contact them.

The maths behind the calculator

For the detail-minded and inquisitive amongst you:

Following standard market research conventions, the formula used is as below. Note that the value in the calculator is the finite size, calculated with the equations below and rounded up to the nearest whole number which is typical for this statistical calculation.

A z-score is the number of standard deviations (e.g. the variation) the value is from the mean. To find the right z-score, you can use the below figures for inputting to the formula:

These statistics form the basis of the calculator input behind the scenes, so you can easily see the output.

That's it! If you have any further questions or queries about the sample size calculator, or want a chat with the Customer Research Team, don’t hesitate to get in touch.

Did this answer your question?