The future of SaaS (partnerships)

I recently gave a talk at Eventbrite about the future of SaaS partnerships with respect to what I've learned building Zapier.

I turned it into a post below, but I think you'll enjoy it more if you watch the video! Choose your own adventure.

Before we can talk about the future, let's go back to the beginning with a scenario.

Imagine you're building a fancy new AI-startup that automagically delivers relevant birthday gifts for your employees when it's their birthday.

You finally launch and get a flood of traffic. Your support channels light up with questions like:

You put all of these ideas onto your product backlog. But the requests don't stop coming.

Naturally you want split the ideas into two buckets: Which ideas are part of our core product and which are integrations with other products?

Congrats! You've just begun the journey towards product market fit. The secret is figuring out where to draw that line, and how the line changes are you grow.

You soon realize you'll never be able to build everything everyone wants.

In the old days, however, you had no other option. If your users wanted a CRM, you built a CRM. If they wanted invoicing, you build invoicing.

But SaaS in the last 10 years has exploded. The best SaaS companies are vertical. I mean that their core product or business primarily does one thing really well. In fact, 7 of the top 10 SaaS companies in 2014 by revenue are in verticals.

Your main core competency is birthday AI software. Are you really going to build a better CRM than a company focussed on it? By definition you will lose – it's not your core expertise.

You'll get nearly all the upside by partnering. A well-known tactic for adding features is to either build or buy. I propose a modern alternative: you should decide to build or partner.

Comparing the two, building is slow and expensive. Partnering is fast and cheap. I'm not saying you should always partner, but it makes sense to in a lot of cases that aren't your core focus.

Over the years at Zapier, here's some common negative feedback I've gotten about partnering:

My users can't figure out another piece of software!


I can build a better experience on my own!

If you're users are asking for an integration, it implies they are using another piece of software.

And perhaps you could build a better experience, but why would you want to? It'll only fragment the already scarce resources you have to begin with.

There's lots of ways you can partner with others.

1. Integrate

This is the most popular method. You leverage existing APIs. In our birthday AI scenario, perhaps you:

The obvious problem is you get too many requests for integrations and you can't build them all. Imagine how much worse it'd be if you weren't only trying to integrate but build all those feature sets, too!

The solution is to pick off the top 1 or 2 most requested integrations, make those core, and rely on other strategies for the long-tail.

2. Platform

You integrate with others using their API, it makes sense to have one yourself. The extension of this is to build a platform where you get to highlight the best integrations that exist.

This might seem like more work than just building the functionality outright, but let me give you some context:

Eventbrite built 5% of their intregrations in-house. 95% were built by partners.

Sure, Eventbrite is a bigger company. Maybe they can command that ratio. Here's another anecdote.

Zapier has 2 engineers spend 4 weeks to build and launch the original platform, launching with 13 apps that weren't built in-house. This all happened 1 week after launching in Techcrunch. Zapier had no brand recognition or clout – it's possibly to built a platform even when you're small.

Let me let you in on a secret: integrations works.

Zapier did a case study with Typeform and found that integrated customers churned at a 66% lesser rate over the lifetime of their account. This isn't unusual. I have anecdotes from a few other platforms that tell a similiar tale in terms of magnitude.

Here's another secret: integrations directories don't work.

Zapier is in a good position to tell, because Zapier is in every app directory. We've tested and optimized our listings. Yet, for any given directory we're in, we get 5x as much traffic from Google searches than the partner's own directory.

I don't say this as a partner of those companies. I'll always take 1x over 0x. But the point is, directories don't work for users. Many mature platform knows this and they're not sure what to do about it. Directories have a lot of problems:

Users quickly learn this and avoid them. This creates a negative feedback loop where the platform gets disregarded by you, your partners, and your users.

You have to curate from the beginning. If you open the flood gates you'll fight an uphill battle to close them. We made this mistake – trust me on this.

Most people on your platform are after one thing: users.

Partners trade their development cycles, time, and energy for the promise that you will bring them new users. If you're not doing that, your partners will care less and your platform begins decaying.

Platforms work but directories don't. How do you resolve this?

3. Embeds

I wish I could say I was proclaiming the future here, but I'm not. Embeds are already part of the modern web:

Blog comments, Social sharing, Customer support, Email newsletters, Tweets, Calendars, Status pages

What do all these have in common?

They aren't in app directories. They are embedded inside someone's core product, right where user's expect to find it.

Why force your users to find your app directory, search through it, decide which app to use, install the app, when what they wanted was just to send some notifications to Slack?

Embed the functionality right inside your product. Yes, you have to play favorites, but you're already doing that if you are curating. It's the ideal trade you can make with a partner. You give them users and they build great product features for you.

Here's the kicker, you don't lose the users you trade! In fact the opposite, it makes them better users! They are more sticky and churn at a significantly lower rate.

Hopefully I've convinced you that partnering with other best in class products is the way to make yours better and retain more of your users.

Keep in mind your product doesnt exist in a vacuum. Whether you're building an AI birthday startup or the next Eventbrite, all your potential partners are dealing with the same struggles that you are. Every single one of them has a long-tail of requests to hook into other products.

At a macro-level, it's economically unviable for everyone to build integrations with everyone else.

Pick the top 1 or 2 that are the most valuable and own those. Leave the rest to your partners.