In an age of quick insight from AI, shrinking budgets, and a decreasing perception of value, it's useful to have some fundamental principles about insight to promote its longevity and value.
‘What is an insight?’ is a question I’ve been asked a few times - by clients, in job interviews, and by my family and friends when I tell them I head up the ‘Insight and Data’ team at Foolproof. It feels like a trick question when it’s asked so simply, but it really isn’t. The question is a simple one, one that deserves a simple answer. I tend to refer to the definition of insight coined by cognitive psychologist Dr. Gary Klein:
“Insight is an unexpected shift in the way we understand things.”
From an academic perspective, research insights can be theoretical and, in the sciences, are drawn from experimental studies conducted over several months. In contrast, industry product research tends to be run over a shorter, and more frequent time period to fit within the product development process. Industry insights go further, and illustrate implications along with recommendations that can be practically applied to drive business success.
All forms of insight should advance our knowledge and understanding of real-world problems and lead to some form of change; a change in perspective, a change of course, or perhaps no change if a product experience works perfectly - which is quite rare since people and technology are ever changing.
When insight falls short
When reading past research, I have come across insights that are not really insights, but are instead facts, observations, or findings. Reports include tables, and intricate graphs with lots of spikes but little to no context or interpretation. Reading and listening to research findings with limited insights can be exhausting and frustrating.
Sometimes a good deal of data is collected and analysed but not properly understood, and the insight is either wrong or misrepresented. An insight should not leave us thinking, ‘Great, and…so what?’.
Here are some examples of poorly written or unfinished insights my team have come across recently:
“Users don't like the product because it's hard to use.”
This is a statement based on customers using a product, one that has presumably undergone usability testing. It doesn’t specify what feature was hard to use or what that means for the customer.
“London to Paris is the most popular travel journey purchased on the website.”
This is an observation based on data trends, but it doesn’t state why or explain the implications.
“X visitors clicked on the [specific button] over the last week.”
This is not a useful insight because there’s no context. X out of what, and how does it compare to other buttons and pages? Why is the past week worth noting? What does this finding mean?
If the reader can’t see the ‘aha’s’, learn something new, or gain a new perspective from the research, there’s likely going to be pushback on the insights generated, along with lots of questions. If more questions are raised than answers, it means more research needs to be done to learn more about the problem space.
It’s also about value, and generating a return on investment (the time and resources spent executing the research). Insights from all research, whether that’s discovery (formative) research, evaluative research, or product analytics data, should help the team learn something new and then do something with it. It should trigger further thinking, and planning, feed into experimentation, inform strategy, and enable the team to learn more about their customers.
What happens when research doesn't provide good insight?
Insights should help inform good product decisions. What happens if insights are not available or actionable? New features may be introduced by an engineer who thought it might be a good idea for themselves and their engineering community. Or products may be developed simply because it’s what the competitors are doing, and they’re making a lot of money.
These implementations may fail because of a lack of insight. For example, the feature was not well understood by the target customers, product engagement was low since the marketing efforts didn’t sell the product features and benefits, or the product wasn’t relevant for the market and just didn’t get the traction. Access to valuable insights reduces the chance of making mistakes like these, and lowers the risk of ill-informed decisions being made.
As experts in data, it's our responsibility to collate data from multiple sources (e.g., triangulation), that are representative and reliable. Data is then cleaned, processed, analysed, and converted into meaningful and actionable insights.
How to define an impactful insight
Insight is data that is analysed, validated, combined, and interpreted:
- It provides context
- It provides meaning that's grounded
- It provides relevance
- It enables us to re-examine pre-defined assumptions
- It can inspire change
When communicated, insights should be framed through the lens of the customer and include:
- The situational context
- The intended outcome
- The restriction, obstacle, or friction
If insights provide us with an unexpected shift, we need to be mindful as consultants in the digital product space, to ensure we can promote this change in view, helping our clients to follow the course of action we’re describing through actionable recommendations.
Rather than presenting cold, contextless facts, we should instead craft qualitative and quantitative data in to framed challenges. That is what promotes a change in perspective.
That is insight.