It’s here! Download our 2021 Holiday Playbook now.

The Difference Between Assumptions & Insights

Jim Babb
Jim Babb

Decisions are difficult. In an era where data is widely accessible and easier to parse, decision-making continues to elude us. This increase in available information presents us with even more nuanced and complex decisions. Increasingly, it’s the case of overwhelming or noisy information that leads to our paralyzed analysis. Whether the decision is simple and low-risk — picking out a candy bar at 7–11, for example — or more important, like choosing the key features in our latest product, we want to make better decisions and increase the likelihood of positive outcomes.

The source of this difficult decision-making is ultimately twofold: it comes from an inherent uncertainty of the future and a lack of trust in our insights. Our distrust is because we tend to confuse assumptions for insights and vice versa. It’s important not to get too tripped up with semantics and over-think our personal and business dealings, but unpacking how we use and trust information can enable us to make better, faster and more confident decisions.

What's the difference between an assumption and insight?

An assumption is your gut instinct. It’s what you believe, feel and accept to be true without supporting proof. Assumptions are models of how the world works and may or may not align with reality. They are driven by conscious and unconscious biases, which at their heart are built from and by past experiences. While experience is certainly valuable, it’s limited no matter how vast. Assumptions should be a starting point to be supported with broader proof as needed.

Insights differ from assumptions. Insights are truths about the nature and inner workings of a given system. They may seem obvious when uncovered, but they can be difficult to find since they’re hidden in plain sight. An insight is supported by points-of-proof that demonstrate the cause-and-effect relationship between elements. In the context of decision-making, an insight is an assumption that has been tested and validated as true.

How do I turn an assumption into an insight?

The amount of proof required to validate an assumption is tied to the risk involved in the specific decision. Risk is a combination of two factors:

The Uncertainty In A System: The future is impossible to fully predict, but a strong insight has the ability to decrease the amount of uncertainty by highlighting how a system currently works and how it’s likely to work in the future. Assumptions, by their nature unproven, do not decrease uncertainty. They can actually increase uncertainty if they are particularly brazen and are the strongest link to your desired outcome.

The Cost Of Failure: If a decision proves itself wrong, how much damage will the outcome cause? Resources, time, loss of profit, etc… are often the costs of failure. Unless insights and assumptions provide a way to hedge a bet, they rarely effect the cost of failure. Insights and assumptions provide no insulation against a negative outcome — in business there are no “backsies”.

How much proof do I need to make my decision?
Decisions can be modeled to reveal the amount of proof necessary to increase the chances of a positive outcome.
Decisions can be modeled to reveal the amount of proof necessary to increase the chances of a positive outcome.
So I don't always need proof?

Assumptions are powerful. Many of the day-to-day decisions we make are low-risk enough that it’s simply faster and cheaper to go with our gut. However, the single biggest pitfall in decision-making is misidentifying something as being lower uncertainty than it actually is and not testing our assumption to solidify an insight.

At the end of the day, when confronted with a decision, you should ask these three questions:

  1. Will the time and resources we spend testing assumptions be worth more than the costs of failure if our assumptions turn out to be inaccurate?
  2. Have we correctly identified the level of uncertainty in our decision’s outcome?
  3. Is there a strong likelihood that we’ll gain valuable insights in testing? Is the chance too low to consider?