The Serendipity Diaries: Solving Problems
My project Serendipity, which launched on Product Hunt in late November has been running for four months now. Serendipity is a professional introduction network of sorts — every month, users receive one email, introducing them to another professional, randomly selected.
And it works! Over the past few months, Serendipity has been connecting real people, in real life. It’s surreal to think that we brought together people in San Francisco to have lunch or coffee:
And someone from Chicago is setting up Skype calls with someone in Paris:
And some pretty cool coincidences have come out of all this:
We’ve also seen how giving people can be to others. Some have made generous offers to help out people they were introduced to, or to connect them to a friend who might be helpful.
Though it’s amazing to read these emails, the past few months have also been a steep learning curve, especially in understanding what people want from Serendipity, and perhaps more importantly, what they don’t. As Steve Blank remarked, no product survives first contact with users. Serendipity, like any other early stage product, is on its way to find product-market fit. I got some things wrong and made many early mistakes (it’s probably worth doing a post on the mistakes alone). But the fun of building things is not just in the first burst of energy during which the Version 1 is made. It’s also in adapting when your initial assumptions are contradicted by what your users want. It’s in paring your product down, tweaking it, adding features, and possibly in the metamorphosis of the product into something completely different. A lot of the fun in making stuff is solving problems.
In this post, I’d like to discuss some of the challenges we’re facing at this point, and some thoughts on how Serendipity can address them.
The One Metric
If it’s true that every startup should track one metric, surely the metric that’s relevant for Serendipity is the obvious one — the percentage of people connecting with each other, based on the introductions.
The corollary to this is that if, over a couple of months, a user doesn’t make good connections through Serendipity, naturally, they get frustrated. As far as they’re concerned, they’re not receiving the promised value. It’s even more disappointing for users when they reach out to their match, but do not get a reply. At best, it makes them reluctant to take the initiative to reach out to their next match, and at worst, it makes them quit.
If this is our most important metric, then improving this metric is what we should focus on to make Serendipity a great product.
Digging to find the reason
To understand what was going on, we ran surveys and collected feedback about whether people ended up connecting with their matches, and if not, the reason why. For the month of January, 41 people responded to our survey. Out of this number, 27 people told us they hadn’t made contact with the person they were intro’d to. (Bear in mind that people who didn’t end up connecting with their match are likelier to give feedback regarding this.)
The reasons were:
Survey results of other months showed similar results.
Relevance of matches
Relevance appears to be a major reason people do not connect. This is an interesting point in and of itself, when you think about how people are using (or want to use) Serendipity. Serendipity is a service to connect random professionals, and during onboarding, users are asked if they’d prefer to meet someone within a particular industry, or if they’d prefer complete randomness.
An interesting piece of statistic is that the vast majority, (nearly 80%) say they want to be connected to someone completely random.
It seems at first glance difficult to reconcile. What do users want from Serendipity — are they looking to use Serendipity as a general tool to bump into interesting professionals, or are they using it with more purpose? Is this a case of people saying one thing (that they want to meet someone completely random) and really wanting another?
Deliberate randomness?
From my talks with several users, I think the sweet spot is somewhere in between. Users want to meet random people, but not so random that there are no common areas of interest. Our job then, will be to maximize this deliberate randomness, the zone where things are random enough to be interesting but relevant enough to be practically useful. To an extent, we’ve been addressing this from the start, restricting Serendipity to only people working in tech, for the moment. But there are other things we can do too.
Getting relevance right
- Location
We introduced location in the second month of Serendipity, and saw good results from that.
2. Profile similarity
Now we’re playing with matching people with similar profiles. This carries with it the risk of ending up matching people who already know each other, but all experimentation is good!
”I got busy”
There are also users who’re actually interested and want to be active, but are busy this week or this month.
We’re soon introducing an option for users to temporarily mark themselves as “Busy.” If they do so, we won’t send them intros till they feel ready to be active again, at which point of time they can toggle their status back to “Active”.
Inactive Users
There are also inactive users, who signed up but subsequently lost interest. These people likely ignore all introduction emails or mark them as spam. They will not have been represented in the surveys, since it’s unlikely that they’d fill out the survey unless they were at all interested in Serendipity. However, they do exist and are possibly a large part of the reason some emails aren’t responded to. For these users, the only solution is to collect feedback, mark them as inactive and stop sending them intros unless they indicate they’d like to be active again.
Other factors
There could be other factors at play. It could just be that it’s difficult to open up a conversation with someone you don’t have much in common. One user made the insightful suggestion that we provide a few conversational icebreakers.
And to conclude
There are tons of startup blogs about hacks that worked for a startup, and quite a few about startup post-mortems. I love both these, but there aren’t too many that detail the challenges that inevitably arise during every stage of building something. (If you’re interested in this, I really like both Tara Reed’s Kollecto blog and Melissa Tsang’s series on her startups.)
I’d like to keep a semi-regular log of the actual challenges Serendipity faces in this journey, and the results — both good and bad — of our experimentation, and share whatever we learn on the way.
The best part about this whole adventure so far has been that when a “click” happens between two people, and they end up connecting, it’s a genuinely rewarding sensation. It makes you feel like you have a hand in something much bigger than yourself — mysterious, even — the end of which nobody can see now, but will pay off in the future.
Jason Fried wrote about the chain of events that led to his publishing an article in the NYT.
If Serendipity sparks even a few such connections, which might lead to something unexpected or beautiful, many years from now, that will have been more than enough.
