Determining Sample Size: What Factors Matter for Brand Awareness Research?
This is the second post in a series of blog entries on how to plan out and execute a brand awareness research project. In the previous post, I shared 8 research design factors to consider when planning a B2B brand awareness research project. This week, I will share some additional insights on determining sample size needs for your inaugural brand awareness survey when you are cost and resource constrained.
Determining sample size needs for an inaugural brand awareness study is a very important input in determining the feasibility and whether or not a brand awareness initiative makes sense for your company at a given time, as sample size is a major input into the cost equation for these types of initiatives. Consequently, marketing managers and brand managers in startups and expansion stage companies oft time get very nervous about making these determinations. They get pulled in opposite directions when thinking about brand awareness measuring initiatives because on one-end they want to have very precise measurements that will accurately measure the effectiveness of their team’s marketing efforts and leave limited amounts of uncertainty, but on the other end they know their marketing budget is cash strapped and they know some of their resources will be more effective if allocated toward other marketing initiatives.
There are many online calculators for determining sample size for a survey that can be used to determine a sample size for an inaugural brand awareness survey. However, these tools are generally designed to estimate sample sizes for studies with either finite populations, limited error flexibility and/or unknown potential outcomes to the study. By ignoring these factors, a sample size estimate has to account for the worst case scenario in order to meet the target accuracy goals. This inadvertently leads to over estimated sample size needs for inaugural brand awareness research projects. I provided a table in my previous blog entry on 8 Research Design Factors to Consider when designing a B2B Brand Awareness Research project that showed general sample size estimates given a standard population size and an unknown awareness level. This table is great for getting a general idea of sample size needs for estimating project costs, but these estimates are still higher than a startup or expansion stage company will generally need in its inaugural brand awareness survey.
To get a more precise estimate of sample size for your study (lower), you should use a more flexible sample size estimating formula. I recommend using the following, as it can be understood for those that are less quantitatively inclined:
- N = number of people in the target population
- P = the lower expected probability in decimal format in the outcome of your survey (i.e. 0.5 for 50-50, 0.3 for 70-30)
- SE = the standard error or minimum accepted error on either die of the resulting measurement from the survey. This is a percentage expressed as a decimal (i.e. 0.03 for 3% and 0.05 for 5%)
- Z = The confidence interval constant. This input determines the confidence that a survey will produce the expected results within a given standard error. The standard confidence intervals used in statistics are 1.6449 for 90% confidence, 1.96 for 95% confidence, and 2.5758 for 99% confidence. For these types of studies 90% to 95% confidence are both acceptable.
Each of the inputs for this equation can roughly be estimated by determining the following 3 factors:
- Total size of potential buyer and/or user population (N)
- Current brand awareness level amongst this population (P)
- Goal for brand awareness growth (Input into SE)
Oft times many of these factors are unknown, but a company can generally guesstimate these factors within a reasonable range to use in determining sample size. To estimate the potential buyer and/or user population (N), companies can turn to secondary market research and or functional associations that serve or license these types of individuals. These estimates can also be validated by estimates calculated through LinkedIn searches across the target geography for a given functional area. Although these estimates are almost always biased downwards as not every individual in this population is active on LinkedIn, or an active or licensed member of an association. These estimates are generally within a reasonable margin of error and as long as you validate the estimate with a second source, you are safe to use that as an input in determining sample size. The formula is not extremely sensitive to sample size.
Estimating your company’s current brand awareness level (P) is a little more difficult without surveying a subset of the population and is a very important component in determining sample size. The lower the expected current brand awareness level, the lower initial sample size that will be required for brand awareness measurements that can be used to accurately test the effectiveness of your brand awareness marketing efforts. Companies can generally develop a ballpark estimate for awareness based on previous outbound calling or marketing efforts and current market penetration. For younger and less developed companies, this is generally 10% or less and this represents a reasonable conservative baseline assumption. However, for younger companies with stronger marketing presences, a more reasonable conservative assumption may be 10% to 25%. In most cases, this assumption is very reasonable and helps you avoid having to consider that brand awareness size could be at 50% which would be at the peak of the spectrum in terms of sample size needs.
The last factor you need to estimate is your company’s minimum targeted brand awareness growth goal between the inaugural survey and the first follow-up survey. Therefore, you first will need to determine the time period between studies, so that you can accurately gauge the brand awareness growth expectations for your team during that period. This information is very important as this will determine your company’s minimum tolerance for error for testing your brand awareness growth goal from period to period.
Using this information, your company’s expected current level of brand awareness and its desired confidence in the measurement results, you can determine the maximum standard error that will enable your company to test its brand awareness growth goal from period to period. To do so, you need to calculate the confidence interval around your estimated current brand awareness level and your minimum brand awareness level to fulfill your minimum growth goal. The formulas to estimate the confidence interval around your estimated results and targeted results are:
Once you have the confidence ranges for each of your results, now you can identify the combination of standard errors for the inaugural study and the follow-up study that make the most sense for the goals of your study. An acceptable combination will be any combination where the current awareness level confidence interval does not overlap with the targeted awareness level confidence interval or zero. Just remember that it makes more sense to have more precise measurements in your follow-up survey because the results of that survey will affect the sample size requirements of two 1st and 2nd measurement periods, not just the 1st. You can easily set-up an excel spreadsheet to evaluate your options in the following format below. Green highlighted entries indicate that there is no confidence interval overlap and red indicate that there is a confidence level overlap.
Using this approach you should be able to identify a standard error estimate (SE) that you are comfortable with for your company’s inaugural brand awareness survey.
Now that you have the 3 inputs into the more flexible sample size estimation formula that I suggested using earlier in this blog post, you can estimate a more precise sample size for your inaugural brand awareness survey. This same procedure can also be used to estimate the sample size requirements for the follow-up rounds of your company’s brand awareness survey. After the first year of your study, you can use the previous year inputs in the formula to determine sample size requirements. The only factor you will need to estimate is the minimum brand awareness growth goal expectations for the following period.
In conclusion, effectively determining sample size needs for a brand awareness research initiative can save your company thousands of dollars in research expenditures spent on unnecessary precision and significant amounts of time in unnecessary data collection.
Leaders from Twilio, IBM, SurveyMonkey and more share their best tips.
B2B brand and product positioning will only continue to become more important with the rise of the End User Era.