The Law of Funnels
No matter what you’re building, there is no way around one of the most important rules of product development: the Law of Funnels.
In the universe of product development and design, most “rules” last just a short while before successful counterexamples emerge. Yet, as with most sciences, there are a handful of fundamental laws that, despite all of our best efforts, are impossible to defy. In my experience, I have come to learn of one such unyielding, ubiquitous, undeniable truth. I call it the Law of Funnels. And it quite simply says that the more steps a person has to go through to do something, the less likely they are to complete it.
It seems self-evident, but many new products and features are built each day without keeping this in mind. It’s hardly a statement that governs only consumer software products. The more chapters a book has, the less likely a reader is to reach the end. The more aisles a person must navigate to find what they want at a store, the less likely they are to buy the thing. Yet this concept is particularly worthy of discussion in the context of product development, because it’s there that the Law of Funnels is so often ignored and so often the lowest hanging way to improve user engagement.
It’s called the Law of Funnels because all users of a product go through a journey that can be thought of as a funnel:
It starts broad at the top with 100% of users (say, people who launch an app). That’s the blank canvas, the point where the sky is the limit and where you have full potential to bring them along for a journey. Yet what one quickly realizes is that as users progress through steps in their journey (i.e. down the funnel), some of them drop off. Each new interaction, however insignificant, requires energy, or input, or action. And so, bit by bit, that 100% shrinks, more and more, until it reaches some number at the end that is almost certainly much less than 100%.
In physics, there is a comparable universal truth. Changes in our physical world—whether something is moving from here to there, transformed, or interacting with something else—suffer from the same law (except it’s called the Second Law of Thermodynamics). Pretty much all physical phenomena (or, in our case, user interactions) give off heat (in this case, users). Lost heat can’t be reused (and similarly, lost users typically can’t be regained). And the usable energy left at the end is less than what we began with.
Consider that customer acquisition cost is inversely correlated to the steepness of your funnel. If only 50% of all users complete your signup flow, then you’re paying twice as much to acquire new registrants as you would if you could reduce the friction of that flow to zero. (Of course, there is no reducing friction to zero. But any conversion above 50% would be a reduction in cost.)
(A quick aside: If you’re enjoying this article, please consider subscribing to my free weekly newsletter, Z-Axis!)
There are a few simple ways to combat the law of funnels by reducing the friction that causes drop-off.
Rule #1: Reduce all user flows to the fewest number of possible actions (but no less than that)
If every action causes a small amount of churn amongst a user base, then the aggregate churn is reduced if there are fewer actions to take.
For instance, builders of user onboarding flows often split the information collected during signup processes into multiple screens. And while this might appear on the surface to be the same number of actions as combining the inputs into one screen, it may not be. Inputting a username and a password on two separate screens in fact requires four actions (two for the inputs and two for the submissions), as opposed to three on a single screen (two for the inputs and one for the submission). While seemingly inconsequential, many products drastically improve their downstream conversion rates by addressing such innocuous-seeming design decisions.
Does that mean all signup information should be collected on a single screen? Definitely not. It’s a balancing act. At a certain point, the overwhelming nature of too many inputs becomes detrimental and more confusing, thus increasing friction, and raising the overall aggregate churn. The number of actions should only be cut down when doing so will reduce the aggregate churn; at a certain point, there may start to be negative returns.
Rule #2: Low friction steps should occur before high friction steps
The mathematically inclined may challenge this, citing the commutative property of multiplication. If three steps each have conversion rates of 90%, 80%, and 50% respectively, isn’t 90% * 80% * 50% (which equals 36%) the same as 50% * 80% * 90% (also 36%)?
Mathematically, yes. But psychologically, no. As users make their way farther down the funnel, they are more likely to become attached to their sunk cost and less likely to be willing to abandon their investment in the product thus far. One should optimize to have as many users go through as many steps as possible before being presented with challenges, on the assumption that more of the users will by then be pot committed.
This rule, in practice, means that an unnecessary step in the process should happen far down the funnel, farther than steps that are strictly necessary. As an example, requests to enable push notifications or to pay for a premium account should be presented after everything else, ideally after some time has passed (i.e. so far down the funnel as to be below regular engagement with the product).
Rule #3: Optimize funnels for the most valuable cohort of users, not the ones most likely to convert
The Law of Funnels manifests itself in different ways for different cohorts of users, causing certain ones to have higher churn than others. Those building products occasionally see groups of users that perform better and assume, erroneously, that their funnels are highly optimized.
Often, the cohort of users who are most likely to successfully go through a funnel are not representative of the highest value users. Perhaps the cohort is made up of power users or early adopters. It is for this reason that averages and metrics can often mislead, and the most effective way to validate and optimize a funnel early on is to test it on a cohort of users that you want but do not yet have.
Bringing users into a funnel in the first place is one of the hardest and most expensive processes for someone trying to grow a product or service. One should do everything they can to hold onto that top-of-the-funnel user base for as long as possible. It’s similar to the platitude about employee retention at a company: it’s significantly more expensive (in time, resources, training, and sometimes cash) to hire someone new than to find ways of holding onto those you have.
P.S. There are many other applications of the Second Law of Thermodynamics to product development, some of which I hope to write about in the future. The Second Law is a fascinating topic, punctuated by the universally held scientific belief that nothing can possibly challenge it. Physicist Arthur Eddington said it best. “If [a theory] is found to be contradicted by observation — well, these experimentalists do bungle things sometimes. But if your theory is found to be against the Second Law of Thermodynamics, I can give you no hope; there is nothing for it but to collapse in deepest humiliation.”
If you enjoyed this article, please consider subscribing to my free weekly newsletter, Z-Axis!
