The more users put into a product — data, content, preferences, connections — the more valuable it becomes to them and the harder it is to leave. Investment is not just retention. It is a mechanism that makes the product genuinely better the more it is used, creating a pull toward continued engagement that grows with every session.
In his 2014 book Hooked, Nir Eyal described the investment phase as the fourth element of his habit-forming product model: Trigger → Action → Variable Reward → Investment. The investment is what the user puts into the product — not money, but time, data, social connections, content, learned behaviours, and customisation. Each investment loads the next trigger, making the next action more likely. The loop closes and starts again, each pass slightly deeper than the last.
The insight is that investment changes the product itself. A Spotify account after three years of listening history, liked songs, and curated playlists is a qualitatively different product from a new Spotify account. The recommendations are better. The interface reflects preferences. The playlists represent hundreds of hours of curation. None of that value exists in a new account. A user who leaves takes none of it with them.
This is the investment loop: the product gets better as users invest in it, which makes users more likely to return, which produces more investment, which makes the product better. The loop is most powerful when the value created by investment is personal, specific, and non-transferable.
“The investment phase is what separates products people use from products people need.”
— Nir Eyal, Hooked, 2014
Discover Weekly launched in 2015 and quickly became Spotify's most successful product feature. The mechanism is an investment loop: every song you play, skip, save, or add to a playlist is a signal. Spotify accumulates those signals and uses them to generate a new 30-song playlist every Monday. The playlist a new user receives in their first week and the one the same user receives after a year of listening are produced by the same algorithm — but they are unrecognisable as the same product.
The two phones below show Discover Weekly for the same user at two points in time. Nothing was filled in, rated, or manually entered. The investment was entirely passive — a by-product of simply listening.
Popular tracks from genres you glanced at. Algorithmically generated from almost no signal.
Niche artists, correct tempo, artists similar to ones you discovered yourself. Uncanny.
The investment that separates these two playlists is entirely passive. The user did not fill in a preference survey. They did not rate songs. They listened, skipped some, let others play twice, saved a few to playlists, and opened the app at specific times of day. Spotify accumulated all of this and used it to produce a Discover Weekly that, by week 52, consistently surfaces artists the user has never heard of but immediately recognises as exactly what they would have chosen themselves. That experience — the feeling of being known by a product — is the payoff of the investment loop.
Spotify published data in 2015 showing that Discover Weekly was streamed by 40 million users in its first year, generating over 5 billion song streams. The retention effect is real: users who engage with Discover Weekly churn at materially lower rates than those who do not — because they have an investment the rest of the platform does not.
The moment a user tries to delete their account is where investment loops become most explicit. Most products show a simple confirmation — “are you sure?” A product that understands investment loops does something different: it shows the user everything they have built, so they can make a fully informed decision about what they are giving up. The investment is still theirs to lose. The product simply makes sure they can see it before they do.
Below is what a well-designed delete account screen looks like for a user who has been using a project management tool for 3 years. Every number is specific to their account. Type DELETE in the field to confirm.
Not a manipulation — a receipt. The screen shows the user exactly what they built, so the decision to leave is informed, not obscured.
This screen is not trying to manipulate. It is not threatening the user or manufacturing urgency. It is showing them their own investment — the 101 projects, 338 workflows, 12 team members, and 36 months of history that they built — so they can make the decision knowing its actual cost. The 73 working days estimate is the switching cost made concrete. The product does not prevent the user from leaving. It ensures they leave with accurate knowledge of what that means.
Most users who reach this screen and see the full picture reconsider. Not because the product scared them, but because the investment loop has been doing its job: the product has genuinely become harder to replace over three years of use. The screen is just the moment that becomes visible.
The investment loop design question is: what does your product learn from being used, and how does that learning make the next use better? If the answer is nothing — if a user who has been using your product for two years gets the same experience as one who signed up yesterday — there is no investment loop and therefore no compounding retention.
Building investment loops means deliberately designing what the product should accumulate, how it should use that accumulation to improve the experience, and how quickly it should return value. The feedback needs to be fast enough that users notice the product getting better, but the investment needs to be deep enough that switching remains costly long after they have stopped actively noticing the improvement.
Eyal, N. (2014). Hooked: How to Build Habit-Forming Products. Portfolio/Penguin. · Spotify (2016). Spotify Discover Weekly: How machine learning finds your new music obsession. Spotify Insights. · Thaler, R. H. (1980). Toward a positive theory of consumer choice. Journal of Economic Behavior & Organization, 1(1).