VarageSale was the most engaging (stickiest) product I ever worked on.
When the company raised $30M+ from Sequoia, Lightspeed and others, one of the investors said (and I’m paraphrasing), “I haven’t seen similar user engagement since WhatsApp.” (which he was also an investor in.)
VarageSale = WhatsApp … well not quite. 😅
But we did have insane engagement.
During my time at VarageSale, we saw ~40% DAU/MAU — which means 40% of our monthly active users (MAU) used the product daily (DAU). That’s very sticky.
Even with great engagement, you have to work hard to maintain it, and to motivate the key behaviours you want to encourage in your product.
That’s when we decided to launch a feature called Top Members.
Top Members was heavily inspired by Airbnb’s Superhost. In fact, back in April 2015, Lenny Rachitsky and I had an email exchange about the Superhost feature (which was in early testing.) Lenny was the product manager for the feature at the time. Here’s a bit of that exchange, which is pretty cool:
Lenny: “The basic story is that the only quantifiable metric it moved is a hosts sense of feeling appreciated (only if they are super hosts). We only ran the experiment for a quarter but we saw no impact on host success, behavior, or loyalty in the data. Qualitatively, hosts love it and are extremely happy when they become super hosts.”
It makes sense that superhosts would love being recognized, but it’s interesting that it didn’t immediately provide additional quantifiable value. If you research into now, you’ll see a number of detailed reports on the benefits of being a superhost.
Lenny: “The biggest win is host happiness, and giving all the other hosts something to strive for if they want to be really into Airbnb.”
Goal setting is critical—you’re giving people something to strive for, which may increase people’s commitment to winning on Airbnb.
Lenny: “…showing a responsiveness number publicly to guests makes hosts obsessed with it. It's a huge lever. Give your sellers a metric to measure their behaviour by and they'll optimize for it.”
Along with the Superhost feature, Airbnb showed seller responsiveness (i.e. “responds in X hours.”) Clearly this made a big difference! We copied this.
I also asked Lenny what % of the user base they wanted as superhosts:
“5%. We originally planned for 20%, but that would have included way too many of our active hosts. 5% seems to be totally the right answer, just on a gut feeling level. It actually just grew to 6% recently.”
Thanks for being there Lenny — I appreciate it. 😄
How sticky is Focused Chaos? Only one way to find out! Please subscribe below.
Top Members may have been the best feature I ever helped develop at VarageSale.
It did what we hoped it would (and more), creating real value in the community.
What was VarageSale?
I still love this logo and font
Before continuing, it’s important to explain VarageSale. I should also point out that while I’m speaking in the past tense, VarageSale still exists. It’s not receiving a ton of attention (as far as I can tell, based on product upgrades) but it’s still there.
VarageSale was a hyper-local buy & sell community focused on safety. We primarily targeted moms who were buying & selling smaller items (often baby-related.) The average transaction price was ~$20. Many of the communities were in middle or lower income neighbourhoods, and the money people were making was meaningful. People sent us stories all the time about the impact VarageSale was having on their lives — allowing them to pay bills, meet real friends & support groups, etc. It felt good to be building something that truly mattered in people’s lives.
Communities were run by admins, who were extremely passionate (which is a good and bad thing!) Think Reddit moderators on overdrive.
And when I say hyper-local, I mean it. Many communities were very small geographically because people didn’t want to, or weren’t able to, travel great distances for small items. In some places where populations were less dense, the geographies grew, but admins were very sensitive about the boundaries of their communities (this was a good and bad thing; can share more in the future.)
I’ve used the word community a lot because that’s what these were. If you compare to something like Nextdoor (which is mostly crazy people ranting about potentially dangerous criminals wandering the streets), VarageSale communities often felt like actual families. (Mind you, Nextdoor won and VarageSale didn’t, so…) We supported our communities actively with a great team of community managers and tried to foster safe, happy and engaging experiences for everyone (we succeeded a lot at this, but it was bloody hard.)
Building Marketplaces
When building a marketplace you need supply and demand. And then you need engagement between the two.
At VarageSale, we started with supply — finding sellers. We knew from trial & error, how many sellers and items we needed to kickstart a community. Once we had that number, we could bring in the buyers and most of our communities would keep growing organically (albeit slowly, because they were hyper-local.)
Perhaps in a future newsletter I’ll share more about how we built supply and demand, but for now, it’s important to recognize three things we needed to build successful communities:
Daily transactions between buyers and sellers
Constant inventory of new items to buy
Trust
We were always thinking about how to increase inventory, drive daily transactions and maintain (and improve) trust within the communities. This is what led to building the Top Members feature.
How Top Members Worked
The feature is exactly what you would expect—a way of identifying and highlighting the best users. Because our communities relied on inventory, we focused on sellers as top members (not buyers.)
Note: Before Top Members, VarageSale already had a “Praises” feature, which allowed you to give someone a kudos for being a great community member (either as a buyer or seller.) We decided not to use a ranking system (i.e. 5 ⭐️) because it felt less personal and it often leads to bad actors (people giving 1 ⭐️ for very little reason.)
To build Top Members, we needed to decide on the criteria. Here’s what we went with (and I don’t think this has changed since we launched the feature):
Need 5 praises minimum (all-time): A baseline that was fairly easy for people to hit. (I can’t remember what % of users had 5 praises all-time though.)
5 transactions per month: The goal was to encourage you to stay active on the platform and sell more stuff.
1 new praise per month: This meant that you needed at least 1 transaction per month with someone new. (You can only praise someone once.)
0 inactive weeks per month: Unless you used the “vacation feature” (which let people know you weren’t around) you had to be engaging (doing something) on the platform weekly.
To determine the criteria we looked at all the user data in our system to identify existing power users that were super engaged. We knew we wanted to encourage more users to become power users (if I remember correctly, part of this meant having at least 100 items for sale), but we didn’t want too many top members. I believe our target was 10-15% of the user base (higher than Airbnb’s original target, because we felt like it was important to make Top Members more achievable and drive a stronger community feel.)
Top Member status was recalculated each month (so you could get it or lose it quite frequently.) And we let you track progress in the app.
I used to be a Top Member :(
Top members received a badge on their profile, which signalled to everyone else that they were a trustworthy & quality person to engage with. When you clicked the gold medal icon it displayed a modal explaining the Top Member program.
Results from Launching Top Members
Admittedly, my memory is a bit fuzzy, but I reached out to a couple people from VarageSale, and together we were able to piece a few things together.
Without question, Top Members worked.
People who received the Top Member badge were thrilled. They loved being recognized for how much they were contributing to their VarageSale communities. We knew people would like being top members, but we didn’t realize how much. People wanted the recognition, they believed they were doing something good by growing their community (not just selling stuff to make money.) Recognition is a powerful emotional force.
Quite a few people worked to get their Top Member badge. We saw a 30% MoM increase in “power users” (again, I believe this was people with 100+ items to sell.) That increased inventory led to increased transactions (although I don’t have a stat on transaction volume increase.)
Complaints dropped ~15%. This was huge! In any social platform like VarageSale, but especially one where there are transactions between people, you’re going to get complaints. We saw our fair share. We invested a ton in developing thoughtful and thorough rules to try and mitigate issues, but they still happened. But when people are gunning to be top members or maintain that status, suddenly they stop complaining. And they behave better in the platform too, receiving less complaints about them as a result.
I don’t recall people complaining about not being top members. The criteria were fair and completely achievable. We didn’t provide top members with any additional benefits—i.e. having their items rank higher, giving them more opportunities to boost their items, etc.—it was purely a visual indicator that someone was active & behaving well.
We struck a good balance between accessibility and “elitism.”
A lot of products (games especially) have levels and achievements. I think of Top Members in a similar way—it was an achievement that people could unlock. You can build this sort of thing into the product from the get-go, or layer in these types of engagement features once you have early traction.
Building engaging products is hard. In fact, I’d say most products fail at this, and don’t create enough value for users, even if they continue to scale their user base. At some point if you don’t have a truly sticky product, you’re going to fail.
Here are 4 steps you can take to build an engaging product:
1. Small loops with fast paths to value
How quickly can you make someone say, “Holy shit, that was amazing!”? ← Figure that out ASAP.
The idea of a loop is something that a user goes through (several steps), resulting in a reward, which further encourages habit-formation and repeat engagement.
Nir Eyal’s booked Hooked does a good job of explaining loops and providing a model for delivering them:
Not every sticky product follows this precise process, but the key is to make sure people get a reward quickly; they need to have an “aha!” moment fast, otherwise they’ll disengage.
Whenever I think about building a product or a feature, I think about loops.
How will people discover the product/feature? What will they be doing before?
What will the product/feature do, and how can I make that as easy as possible?
What benefit/reward will they get?
Will that benefit/reward be sufficient enough to keep them engaged/doing more? (Possibly in exchange for the value giving me something in return; i.e. attention, data or dollars?)
How quickly can I get them back doing the same thing (or a variant; expanded version), and ensure they have the same quality experience?
You can think of this as a customer journey as well (and I don’t think startups do a good job of truly understanding customer journeys!) Often customer journeys are thought of in a linear way—instead think about loops.
Ben’s B2B Software Pet Peeves
Long-ass tutorials: Don’t do this. It sucks. Everyone just clicks through everything to get into the product.
Empty screens: There’s nothing more disheartening than an empty dashboard. Unfortunately most B2B products try and resolve this with long-ass tutorials (but you know those suck.) Figure out how to break up your product into smaller loops, and guide people through those. That way you can expose key components of the product as need be, and ensure that the default isn’t blank.
2. Define and understand the expected frequency of use
Every product has an ideal frequency of use that demonstrates users are getting value. Some products are daily use, others you use infrequently but they’re still valuable.
You need to define the expected use for your product/service and then build functionality that drives towards those usage patterns.
If you’re building a social app, chances are you need daily use to prove stickiness and have a chance at building an actual business with a viable business model (i.e. Instagram, Slack, Snapchat, Discord, etc.)
If you’re building a B2B app, I’d aim for weekly use (monthly at the most)—anything less frequent than that and I can guarantee you people are questioning their spend. Some B2B apps might operate behind the scenes, but they still remind you that they’re there and creating value at a regular cadence (i.e. Cloudflare, which you may set up for a website and not regularly check, but they send you updates so you know it’s working for you.)
The challenge with products/services that are meant to be infrequently used is that the testing and learning cycles are slow. Think about tax planning software, which may get heavy use during tax season and sit abandoned otherwise—if you miss the mark, you’re waiting a long time to fix things.
Defining ideal usage patterns should impact what you build; because you’ll focus on features that generate the type of usage you believe drives value. Knowing those patterns and the baseline of behaviour also lets you figure out what metrics to track and benchmarks you’re aiming to achieve and improve on.
Sometimes you get the target wrong, but that’s OK
In 2011-2012 I invested in a startup called HighScore House through Year One Labs. Their pitch was to help parents get kids to do their chores. Parents would put in weekly tasks, kids would do the tasks and then receive points that they could redeem for rewards.
We challenged the founders to define their expected usage patterns—how often were they expecting parents and kids to use the platform? By setting a measurable benchmark, we could see how well they were doing, and make better decisions.
If I remember correctly, the team said, “We expect parents and kids to use HighScore House 4 times/week.” The expectation was that parents would set the tasks at the beginning of the week, and kids would be doing things all week, and coming back ~3 times or so per week to cross tasks off.
Unfortunately, that’s not what happened.
Usage was significantly lower than that; in fact, most kids were only using HighScore House once per week. Parents were behaving largely as expected, but kids—no go.
The team fretted for awhile, and iterated a bunch to try and increase usage. And then the CEO, Kyle Seaman, did the smart thing and he called the parents.
Amazingly, he was getting a lot of positive feedback. Parents were telling him that their kids were actually doing their chores, and they loved the product! So Kyle then asked the obvious question, “But your daughter is only using HighScore House once per week, are you sure it’s working?”
The response was (in hindsight) obvious, as a number of parents would say, “Kyle, we are trying really hard to limit her computer time, so she does her chores throughout the week, and on Sundays we give her a couple hours on the iPad. That’s when she goes in and crosses off tasks, gets her points and enjoys the app.”
Ah! Parental limits. Duh.
In this particular case, with new evidence/data helping us understand how people were using HighScore House and getting value, we were able to “lower the target benchmark” — suddenly the % of active users jumped up (which I know feels like cheating, but it’s not) — and it increased our confidence that the company had built a valuable product.
3. Identify your best users and learn more about them
Assuming you have users, there’s a very good chance some of them are “crazy” (but in a good way!) These are users that use your product way more than you’d expect, and ideally they rave about it (to you and others.) If you don’t have any crazy users, you’re in real trouble.
Instead of “crazy” let’s call them your “best users.”
You need to understand your best users intimately.
What makes them tick? Why are they behaving differently from everyone else? Are there similarities between them?
Figure out the commonalities between your best users as quickly as you can through quantitative and qualitative data; i.e. measure stuff and talk to them.
Those commonalities may be the path towards winning. If you know the similarities between your best users, you might be able to:
Figure out how to find more of them (and in turn attract a higher quality user)
Adapt your core value proposition in-market to attract similar people looking for similar things
Enhance specific functionality based on what they love about your product (“Give me more of that stuff!”)
Experiment with ways to turn your “less good users” into your “best users” - because you’ll know the thresholds of usage that drive maximum value creation
Knowing who your best users are, and figuring out why they love your product unlocks pretty much everything and has a massive impact on your go-to-market strategy, roadmap, etc.
4. Reward your top users
Look how happy they are!
And this all comes back to Top Members and Superhosts. 😉
Once you’ve identified the best users, you want to strengthen that bond. It’ll ensure they don’t abandon, but also can drive them to instigate more referrals, promote you more, etc.
Reward your best users with perks and benefits that make them stand out. Not only will it lock them in, it most likely will encourage others to want to achieve the status those top users are getting. This is what happened when we launched Top Members—many more users wanted to achieve that level of status.
Often, founders ignore their top users—not because they don’t care about them, but because they think they’re secure. You might assume those users won’t abandon, or that they’re already generating enough value for you, but you could be wrong. Instead, double down on your best users, learn from them, and use that to: (1) keep them; (2) make them even happier; and (3) figure out how to find more of them.
Long live the crazies.
Will you be one of my crazies? Subscribe below, and let’s find out together. Thank you!
Great article! Thank you. :)
Does the Top Member / Superhost concept translate to B2B applications at all? eg. Hubspot
This is a great post! We're building a marketplace right now (www.booomerang.io) and these product tips are handy! Thanks for sharing