Ensuring statistically significant data

When we are developing a pricing strategy, we collect relevant and statistically significant data from your potential customers. If you are conducting research (of any kind) you will need your results to be statistically significant. If you cannot ensure statistically significant results than you cannot generalise from your collected data. This is true no matter what research you conduct, including willingness to pay and setting correct prices.

What does statistically significant mean?

For something to be statistically significant this means that the research sample is large enough receive relevant information. If you are interested in knowing what your customers value in your product, you conduct the data collection. But if you only ask 10 people and give them 8 options, can you actually draw any conclusions from it? No. This sample is far too small to draw any conclusions from, you cannot generalize from the data.

It doesn’t only matter how many people you ask, but also who you ask. Imagine if these 10 people are not even customers or potential customers. Then it would be silly to make business decisions made of their responses. This is why sampling matters. This is why you need to qualify your respondents and choose your sampling properly, which you do through research design.

When we are looking to statistical significant it is the probability we are interested in. How likely, or probable, is it that our collected data is accurate.

How many data points are needed for statistical significance?

It depends. Of course, this is not the answer you are looking for, but it does depend. You may have heard or read that “A minimum of 30 observations is sufficient to conduct significant statistics.”. You can draw conclusions from as little as 30 observations, but then you need to adjust your research design to fit a smaller sample. The algorithms and research design we use at Atenga Insights allows you to get indications from as little as 15 respondents. As is highlighted in the image below (where respondents is referred to as samples). Where you can see 15 samples as the green line, illustrating the curve overlapping to a large extent with the higher number of respondents.

Of course indication is different than conclusion, but sometimes indications are enough to make general business decisions. But for statistical significant you see that above 25 you have a pretty clear view of the overall outcome.

Trusting the pricing recommendations from Atenga

At Atenga we have collected and analyzed data for years. During this time we have refined our research design, this means how we gathered the data. To ensure that the pricing recommendations we set are statistically significant and relevant for our clients.