Creating Product Stickiness

How should we think about competitive moats for companies like Spotify that aren’t quite marketplaces yet still have data and multiple parties involved? Is defensibility (or lack thereof) inherent to the market structure, or can you create stickiness out of thin air? Is there another category of defensive strategy yet to be fully explored?

For Spotify, the traditional network effects model doesn’t quite work as it would for Facebook or Twitch. Sure, you have consumers on one end that incentivize labels to add more content to the platform. You can also add original content, or build things like playlists that are mildly high-friction to transfer.

But labels can raise prices until they capture most of the value. Competing platforms can offer the same content at a different price. And switching-cost fundamentally doesn’t get stronger with scale. In fact, product friction often incurs a tradeoff between usability and stickiness.

Spotify ultimately leans heavily on product design, branding, loyalty, and other forms of “weak moats”.

But I think there’s a whole set of stickiness strategies that companies big and small can exploit to raise switching costs and slowly eat more and more of the user stack.

Example 1: “blank slating”

We’ll stay with music streaming products for now.

The data you generate when playing music can only be so useful: Spotify can personalize playlists to fit your tastes, but Apple Music can easily do so too, especially if you import your playlists. Both companies have large data sets to make recommendation engines good. As much as you think you’re a musical snowflake, most people’s tastes aren’t actually that unique.

But what Apple Music will never have is the list of songs you have saved but have not listened to in a while. So Spotify can always hand you a radio station full of “favorites you haven’t listened to since last year” that perpetually surfaces novel yet nostalgic content.

This relies on data that cannot carry over between competing products. It’s a tree rooted within the bounds of the walled garden. If you switch services, that entire set of data insights becomes a blank slate.

Example 2: “stored data”

Literally every talking point I’ve heard about Superhuman relates to 1) speed, 2) keyboard shortcuts, or 3) the “luxury software” effect.

You may assume that I’m another VC Superhuman evangelist, but I’m surprisingly lukewarm user. Many apps like Spark have plenty of keyboard shortcuts and add-on features. I find myself spamming the ESC key more than I should, and I really wish I could have all accounts in one view like I can with other apps.

But this post isn’t about the pros and cons of different email clients. It’s about untangling which features actually contribute to a business model that’s unbeatable in the long-run.

The most salient elements of Superhuman’s moat are the ease of use, favorable aesthetic/UI, counterparty profile panel, and handy features like read receipts, reminders, and send later.

Unfortunately speed, aesthetic, and simple features are not very defensible. Spark is already ahead on half these dimensions, plus they have team functionality.

What is defensible, though, is the long list of emails you’d like to be reminded about and the emails you currently have queued to send at a future date. Followups and executing planned outbound are both very important.

“Stored data” must be painstakingly carried over to another service. Or slowly transitioned over time. In either case, the user just doesn’t want to make a switch.

If I were on the product team at Superhuman, I’d be doing everything I could to make Reminders and Send Later first-class citizens. Both in the typical email use-case, but also in the “Note-to-self On Steroids” use case.

The beauty of this stickiness-driver is that it’s pulled out of thin air. There’s no trade-off to be made.

(At this point you might be thinking “wait, what about the exclusivity? The invite system! The elusive prestige of a private product!” Sure. The status/signaling element is valuable right now. Just as it was for Facebook when FB only existed at Ivy League schools. But that sort of moat is more of a growth-hack than a form of long-term defensibility. Nobody thinks FB is elite or exclusive anymore. It’s the opposite.)

Lessons learned

I’ve had a tough time coming up with original examples above. At the concept’s core: there are sometimes small niches where you can increase product stickiness without making a friction/usability tradeoff. That’s what contributes to defensibility where there is otherwise not much.

Facebook might inherently be a network-effects driven social product. Spotify and Superhuman have to work a little bit for it. But it can be built!

Whether it’s the “blank slating” approach, the “stored data” approach, or something else entirely, your job as a founder or PM is to find subtle ways to create product-driven stickiness out of thin air.