The insight
How do you know if your insight is any good? What if it’s too obvious? Too obscure? What if your personal bias is too in play? All fair questions.
In the decade+ of mining data at Twitter to find rich insights, we learned a lot about the process. From that, here are some tips and ways to test your insights.
Fill your mind with reference points. Read. Watch. Listen. Our jobs are to join the dots and see connections where others see disparate datapoints. It’s when you can spot something over there that applies over here that you know there is a pattern - and patterns mean meaning.
Search, search, search. Search google, search Twitter, search any place with a search bar. You are not the first to research the topic at hand. Your job is not to say something new, but to say something in a new way. Or to say something, joined with some other lens, that makes our understanding of it improved.
If it’s not succinct, you’re not there yet. If the insight needs a lot of context building or justification, it’s not strong enough. If you have to do a whole set-up, it can’t stand alone. It might need more polish, or you’re not there yet. Sleep on it and try again tomorrow.
If the insight doesn’t set up a tension that your service / product can’t uniquely resolve - it’s not working. If the payoff is tenuous, people will sense it (either when you’re trying to sell it in, or when the work goes live). We sometimes fall so much in love with our own thinking that we forget the payoff needs to work. That might mean walking away from a hard fought insight, or reserving it for another problem.
