How we 10x’d our activation rate by recreating concierge magic in self-serve

When we initially launched Hoop, we onboarded users individually into the new product via a 30 minute Zoom call. It was awesome to get to know early customers, but it was also a bottleneck for growth. As our user base grew, so did the time we spent hand-holding every single person. Not exactly scalable!

Fast-forward a few months, and we’d designed Hoop’s self-serve experience so effectively that our activation rate shot up by a factor of ten. Here’s the behind-the-scenes play-by-play of how we went from concierge to self-serve without losing (and actually gaining!):

Starting with concierge

Hoop is an AI task manager that automatically captures tasks across Slack, email, and meetings. Our goal is to make sure busy professionals never miss another task when going about their busy workdays.

We spent months in the “concierge” phase of onboarding new users, manually onboarding early customers one at a time. On each 30-minute Zoom call included, we set up a new user's account, connected tools (Slack, email, etc.), and walked through the product. We also got to answer questions in real time and objection handle.

This high-touch approach was fantastic for validating our product. We especially learned that showing our product in action by demonstrating task capture was key to getting users to the magical "aha" moment where they realized how task capture worked. We would show this by saying a task out loud like, "Can you please send me your helicopter pilot's license after the call?" after the new user created their account. Their face would light up in understanding when they saw that task pop up in Hoop almost immediately. They now understood that Hoop would auto capture tasks and put them in one spot.

This high touch experience would obviously not scale. While other companies, notably Superhuman, scaled via concierge onboardings, we felt confident we could figure out a self serve flow that matched or exceeded the live call. No busy person wants to be on another call, right?

Then the bottom fell out: When we initially launched self-serve, our activation rate (the percentage of new users who truly got to the “aha” moment) plummeted from our concierge heights down to low single digits. The same magic that wowed people on Zoom calls wasn’t happening by itself in an online signup form.

Defining our north star metric

We decided to take a step back and define a what we meant by “activation.” After exploring a bunch of data and user patterns, we landed on:

A user is ‘activated’ if they receive tasks from at least two different sources within their first eight hours.

For Hoop, these “sources” are Slack, email, and meetings. The moment a new user sees tasks from multiple places, and realizes they did nothing special to get them, it signals we’ve delivered our core value.

Here’s why we picked it:

It captures our core promise: Hoop automatically centralizes tasks from Slack, email, and meetings. No manual input.
It’s time-bound: We wanted users to feel that magic fast, before their initial enthusiasm wore off.
It correlates with real product usage: Our data showed that once people see tasks from multiple sources, they’re way more likely to subscribe and become long-term users.

The three key changes that moved the needle

After launching a basic self-serve flow, we found ourselves stuck in single-digit activation. Worse, we’d lost the personal “Ta-da!” moment from concierge calls.

- We had to recreate that moment asynchronously. We realized we could do a look back and have tasks appear in their Hoop once they onboarded. Once someone connects Slack, Hoop scans recent messages to find real tasks. We cap it at around three tasks so it doesn’t overwhelm anyone. Similarly, once they connect their Gmail inbox, we do a short historical look-back. Within minutes, users can see real tasks from their emails.

- In our high-touch calls, people never believed Hoop would automatically capture tasks, until they saw it live. So we embedded a 40-second video in the onboarding flow, where Justin literally says, “Hey, can you send me your helicopter license by the end of the day?” After a short pause, that exact task pops up in their Hoop to-do list. Yes, we worried some users might dislike an unskippable video, but it quickly recreates the “aha” moment. They see a real example of Hoop capturing tasks on the fly, and it feels live, the same way our Zoom demos did.

- Finally, during onboarding, some users would skip connecting Slack. Initially, we didn't know why. So we added a quick “Why are you skipping?” prompt. The answers gave us immediate clarity:

Security concerns: They worried about data privacy. So we added in-app reassurance and FAQ answers about how Hoop stores data.
Permission issues: Some folks needed permission from their IT admin. We started emailing them instructions to share with IT.
Unclear value: We rewrote certain steps to emphasize why connecting Slack or Email unleashes Hoop’s autopilot benefits.

After incrementally implementing these changes, the difference was stunning: Activation soared to roughly 10 times the early self-serve baseline. People who saw tasks from two sources within hours also turned into our most engaged users, and ultimate subscribers.

For me, the best part is hearing that new signups don’t believe Hoop will truly capture tasks automatically until they see that Slack or email look-back. Or watch that short video and get the helicopter-pilot-lives moment. That’s the same skepticism we used to battle in concierge calls, which we can now resolve at scale.

Takeaways for other startups

Take time to define the right metric- Don’t just pick something arbitrary like “user logs in.” We hammered out an activation metric that directly maps to our product’s core promise: tasks from multiple sources within 8 hours.
Do the concierge (that doesn’t scale)- We ran hundreds of manual onboardings. Yes, it was time-consuming, but it yielded priceless insights into the exact moment people light up with excitement. Those insights drove the entire self-serve redesign.
Recreate the “aha” moment- If your product’s magic is typically delivered in a live meeting or white-glove setup, ask yourself how you can simulate or compress that magic in an async flow. For us, short “look-backs” + a 40-second video were game-changers.
Address objections in real-time- If users skip or disconnect part of your onboarding, collect that feedback right away. The top reasons will shape your copy, your help docs, or your entire approach.
Iterate, measure, repeat- We didn’t fix activation overnight. It took a series of small experiments, watching the data each time, to climb from single-digit to 10x higher. Activation success is rarely about one giant feature; it’s usually the sum of many smaller improvements.

What’s next?

We’re still discovering better ways to onboard folks into Hoop. But for now, the leap from “Hey, this might be neat” to “Wow, I see my actual tasks already, how did that happen?” is night and day. And it’s a direct outcome of spending months in the trenches with our earliest users. We learned what made them go, “Oh my gosh, that’s magic!” and then we productized that magic.

“You have to imagine: we never would have come up with that video or Slack look-back if we hadn’t done concierge onboarding first.” – Justin

I couldn’t agree more. If you’re wrestling with your own activation or onboarding flows, I hope our story helps. Find those little sparks from your high-touch experiences and scale them up in a way that’s immediate, fun, and undeniably useful. Thanks for reading and here’s to making the “aha” moment happen for every new user (without all the manual 1:1 calls)!

Ready to try Hoop for yourself?‍