The problem with growth loops

A growth loop describes a product engagement pattern that engenders the discovery of the product by new users. Ultimately, a growth loop should articulate the mechanics of an audience aggregation flywheel: a user does some thing, that thing is discovered by non-users of the product, those people join the product, and those new users do that thing. A Google images search for the phrase “growth loop” yields a number of examples of growth loop diagrams for various products. My interpretation of the canonical growth loop prototype is below:

The concept of growth loops has been popularized over the past few years as part of the “product growth” school of thought that promotes viral product design. Advocates of growth loops posit that triggers and mechanisms can be architected into a product such that users are compelled to share that product socially in a way that stimulates sustainable, compounding user base growth.

Virality is not a new concept in social product design, and it has certainly been explored exhaustively on this website. But what differentiates growth loops — and, especially, the retrospective, third-party analysis of growth loops — from the measurement and assessment of virality is that growth loops aim to be explanatory. A growth loop proclaims a causal relationship between some interconnected system of product features and the growth of the product, within the specific scope of user-product interaction. A study of virality simply aims to quantify the contribution to a user base of viral discovery.

Last year, I wrote a short Twitter thread about what I see as the philosophical problem with the concept of growth loops:

It’s certainly useful and instructive to dismantle successful products in an attempt to find replicable, generalizable design patterns. But what I find problematic about the study of growth loops is that it sidelines other sources of traffic and places primacy on deliberate, conscientious product design for user base growth. Charitably, this feels reductive; cynically, it feels like a proclamation that the “if you build it, they will come” ideal merely requires a few thoughtful product tweaks before kickstarting a flywheel.

Yes, thoughtful design can compel users to promote a product, either via purposeful advocacy or through natural, perfunctory artifacts of usage, eg. Spotify’s partnership with Facebook that exposed song interaction on users’ Facebook feeds. A user base can grow and compound as a result of these mechanics. But most growth loop analysis, and the subsequent implementations that it inspires, ignores outside influences on growth — the product must have grown exclusively as a result of these mechanics. This is a post hoc rationalization that oftentimes discounts the impact of other forces or product realities:

  1. First mover advantage. Innovative, novel products that are gaining nascent traction with early adopters tend to receive a disproportionate amount of media coverage. This kind of awareness can persist and snowball of its own accord; it doesn’t simply evaporate over time. Any analysis of a company’s path to success can’t ignore that it received a boost off of the starting block from earned media;
  2. Network effects. The ambition of any social product is to seed network effects (“the product gets better as more people use it”), which themselves clear a path to organic discovery. When a product’s network effects begin to blossom, any explanation of some specific, prescribed path that users take within a product to ultimately promote it to other people is moot: organic / viral discovery has manifested into a reality. The particular tool that a user utilizes to evangelize a product is less meaningful than their motivation to do so;
  3. Ongoing paid media. A growth loop analysis explaining the success of TikTok might explore its discovery algorithm, viral challenges, the ability of users to share their TikTok videos in their other social media streams, etc. But if that growth loop analysis ignores the fact that ByteDance spent $1BN on paid advertising to promote TikTok in the US alone in 2018, then it is incomplete;
  4. Retention. The most effective growth mechanic is retention: strong core, long-term product retention increases the efficiency of paid marketing and allows for faster cohort compounding, creating more mass from which a gravitational pull of organic and viral discovery can manifest. The best way to increase the size of product’s user base is to make it sticky and retentive. But the pernicious thing about the growth loop mentality is that it ignores negative condition of what growth might have looked like without a prescribed, deliberate growth loop in place. I have seen cases where the growth of an app that doesn’t naturally produce network effects has been impaired by awkward social features and viral mechanics that weren’t aligned with the product’s core use case. Any implementation of product-first growth needs to be consistent with the product. A deliberately configured growth loop can actually inhibit growth for products that don’t naturally lend themselves to the Three Cs of product engagement that drive social discovery.

The problem with growth loops is that they can premise an unjustifiable supposition — that a successful product grew as a result of a deliberately engineered product mechanic that produced a virtuous cycle of organic product adoption — and work backwards to that justification while ignoring other sources of growth.

Studying successful products is a worthwhile exercise and, especially for social products, great artists steal. But product growth is more often than not the result of an ensemble of efforts. Any analysis of a product’s growth machinations should avoid the confirmation bias of attributing success entirely to one thing.

Photo by Yarenci Hdz on Unsplash