When you are doing your initial customer development, by talking to many potential users, there are many cognitive biases you need to be aware of.
Cognitive biases are tendencies to think in certain ways that can lead to systematic deviations from a standard of rationality or good judgment.
Usually most founders tend to solve problems they have exposure to or those they are aware of, or those they believe to be one that’s a large market. This stems from the “scratch your own itch” phenomenon.
I had a conversation with a founder who is building a consumer internet company, where viral effects of her product determine the growth trajectory more than any other metric. Or so, she had learned from many other founders experiences – both by talking to them and investors in the space.
After 3 months of building her mobile eCommerce product, she and her cofounder launched it in the marketplace. Initial traction was good and trending ahead of their expectations. Many of the early users were impressed with their product selection and merchandise.
Growth after the 4th month though, stalled as they were on the road trying to raise their initial funding. Most every entrepreneur knows that fund raising can be a full time job. In fact I have mentioned several times that fundraising is a poker game more than chess.
When they were trying to show their initial user growth, many investors had the same problem – was their product a trendy, 3-month-uptick or a sustainable-fast-growth business?
After hearing this from the 5th seed investor, they determined that they need to look closer at their numbers, their repeat purchase behavior and address the issue before they were going to raise any funding.
Looking at the initial numbers suggested their they had many buyers who got to know about them through word-of-mouth, and the repeat purchase was high.
She and her cofounder determined that they had to improve their virality coefficient.
This is the bias I see most often: clustering illusion.
The clustering illusion is the tendency to erroneously consider the inevitable “streaks” or “clusters” arising in small samples from random distributions to be statistically significant.
When you have very little data, you have very little data. That’s it.
Don’t make assumptions about the overall market based on very little data.
There are times when you have 60% of the data and you have to make a decision. There are times when you have 30% data and you have to make a decision.
The difference between 30% and 60% is a lot. In fact, most entrepreneurs I deal with confuse having 3% of data with 30% of data.
To reduce clustering illusion the only remedy is to get more data. You will have to run more, smaller, experiments, over smaller periods of time and do it consistently. Make your assumptions, document your hypothesis, but continue to work on getting more data.
Turns out the real problem for our entrepreneur was that the overall market was much smaller, and they found it after 1 year of trying to increase their virality coefficient. They did raise their initial funding, but have since pivoted to expand their merchandise offerings to cater to a larger market.