For many of us -- those of us with products which enjoy the benefits of "network effects" -- one major measure of success for our startups will be the number of users we can attract.
Of course, we know that we're not just looking for a Registered Users metric (RUs). Many sites have millions of RUs who never or almost never return. So, in general, we don't just want millions of Registered User on our services but millions of Active Users using our service. In Startup Land, that's the ball on which we always seem to have our eyes.
Unfortunately, I think all this thinking about millions of users has us a tad distracted (not to mention sounding Dr. Evil-esque. Say it with me... "Millions...").
The problem with thinking about millions of users is that it's so non-specific and directionless. How do you get to millions of active users? Where do you start?
Getting to millions of users, it seems, is not done through press and it's certainly not done "targeting" our sites with the right User Experience and then waiting for the Internet of bees to find and pick our beautiful flower over other beautiful flowers. Too many "great products" never get used.
In fact, I propose the best way of getting to million active users is by:
1. Creating a Product for Very Specific People. Names, people! Names!
2. Hand-picking and Delivering Product for the First 100 people.
3. Focusing on the "Next Logical Node"
Focusing on Specific -- Really Specific -- People
All too often, I hear people talk about user groups they want to "target" in ways that sound like they're buying ads on Google.
"We're really building this for College Students"; "This is designed for the mid-western Mom type"; "This is something early adopters are really going to like"; "Vegan/Urban Jet Setters"; these are all examples (well, except for that last one, that's a joke).
The worst of the worst, however, is: "Normal people."
Normal people!? Whew, I'm getting worked up just remembering the last person to say that to me.
How about names, people? Names! Starting with your own.
When we build products for groups of people we aim for a fictitious middle. While we're trying to make something people LOVE we make assumption about what they love. It would be like saying to your girlfriend "I hear women LOVE Victoria's Secret. Go try this on." when in actuality some women might like that but she really loves sweat pants; taken as an individual, demographics mean jack.
When we build products for real people -- starting with ourselves -- however, we know exactly what to build.
Before he launched Foursquare, Dennis Crowley would go around saying "I just want a Leaderboard for my Saturday night and I want to be at the top of it." Foursquare wasn't just built for himself, it was built with a very specific set of friends in mind -- those he spent his Saturday night's with.
Right now, your demographically-trained mind might be thinking that "people you spend Saturday night with" is a new target market. What were you doing with "Mid-Western Moms" when you could be focusing on "People you spend Saturday night with?!"
But in reality, "people Dennis spent Saturday nights with" was a very, very specific group of people. In fact, they have names: Naveen, Will McD, Kevin K, his brother J, Mike H, Noah B, Michael G, Chad S, Rex S, Rick W, me sometimes... If you go back over why and how Dennis and Naveen built Foursquare, you'll realize that they built specifically with Dennis, Naveen, Will McD, Kevin, J, Chad and the handful of other first users (there were about 20) in mind.
When I talk to folks doing startups -- and when I've built my own products in the past -- I never, ever, ever, ever hear people with this level of specificity. Yet, it's this level of specificity who can help you build the best product you'll ever build.
As it turns out -- again with Foursquare as an example -- they didn't set out to build something for "mobile urbanites" (or, as one person recently assumed, "people who like earning badges" -- ugg, really?), but for real people. And so it was a real person who said the Feed should be on the first tab of the iPhone app and it was a real person (Chad S, as the legend goes) who suggested he wanted to be crowned "Mayor" of a place (Think Coffee!).
Dennis and Naveen knew that if those first 20 people -- the real people with names who they spent their Saturday nights with -- didn't LOVE AND OBSESS over their product, then no amount of tech press, "main-stream" press, user outreach, business development, VC-related PR (really?), "pivoting", focus groups, or redesigns would somehow make their product magically more loved or more obsessed over (as if it wasn't the product that was missing the mark, but their batch of users they splattered at it).
Dennis and Naveen did what you and I should have been doing this whole time (and I've been trying to do with Ohours ): building for real people. People with names. People we know and who will honestly tell us why they're not totally in love and don't obsess over what we've built (because believe me, we'll know when we have). Most of all, building for people with names that start with our own.
The First Hundred People
If you're doing it right, you've now built something two people CAN'T LIVE WITHOUT. (Hint: One of those people is you.)
Next, go focus on getting 100 users. But which 100 users?
I've come to believe that nothing matters more than "who" are the first 100 users of a service.
Now, when I say "who" you're probably thinking "how many followers they have on Twitter" -- influencers, right?
There are two important factors which determine the first 100 people of your service:
1. How much they love your product
2. How many other people on the service they have close relationships with
Back to Foursquare. When Foursquare launched at SXSW two years ago, there were a bit less than 100 people on the service at the time of their launch. Same goes for Gowalla. (100 is a relevant number because it was -- I don't know if it still is -- the number of "Provisions" you could have for a beta version of an iPhone application.)
While their feature sets and marketing spends were different (remember how Gowalla bought out the back page of the conference guide?) the main difference between Gowalla and Foursquare was who were and -- specifically -- where from, their first ~100 users.
From afar, it seemed that Gowalla did what most of us try to do: get 100 "influencers" to use our service. If they love it, they'll tell their friends, and then we'll win. Right? When I got to SXSW that year, Big Blogger So-and-So from SF had Gowalla installed, Super Angel MegaStar had Gowalla installed, my good friends Gary and AJ Vaynerchuk had Gowalla installed, and of course a batch of Austinites had the app installed.
The problem? Few of these people knew and LOVED each other, so it made it extremely hard to know and LOVE the product.
When it came down to it, Foursquare made the biggest splash because it came with a wave of mostly non-influencer-yet-passionate users who all knew and loved each other, and therefore all knew, loved -- and evangelized the shit out of -- Foursquare.
(Note: Dennis points out to me that Foursquare also had an advantage in that 1) he had done it before with Dodgeball and so made less mistakes and 2) that Dodgeball, while it only had a few passionate users still remaining, was shut down by Google only a month before SXSW, so those stalwarts were inclined to join the new product.)
Whether they meant to or not, Foursquare realized that network and usage density trumped external "influence."
Out of the early provisioning codes, I more than 90% of them were given to NYers. Why? Because with their initial user-base, it was more about who you knew already on the service than who you knew -- or how many people knew you -- off the service.
At the end of the day, I posit that the network density of the first ~100 users of these services, above almost all else, defined their future trajectory. I posit that when we build services, we should care about "who" we have as users; but "who" as defined within the service, not outside.
Next Logical Node
Hopefully, we're now building products for very specific people and thinking that quality of the network is defined by the relationships members have within it, not the influence and relationships outside it.
But, more specifically, how do we go from 2 amazingly passionate users to 100? Once we've hit 100, how do we target the next set of users?
To answer this question, I've come to believe that there's always a "Next Logical Node" in the network.
To illustrate, let's say I've built something I can't live without (I have! It's called Ohours!) and I have one friend who also loves the service and can't get enough of it (I do! His name is Vinny!), how should I go about targeting new users for the service such that the service has the best chance of being a huge success?
I can't tell you how many times people have pointed out to me that I "need to get Fred Wilson" to use Ohours. It's obvious, right? Fred would be a ideal next node in the network because a lot of potential hosts and attendees look up to him and may follow suit? Right?
Fred would bring in a lot of hits and signups, for sure, but a) without a base of other hosts of the same variety (i.e. ultra-high demand) most potential attendees would vie for one of his limited availability and then leave the site, leaving a very poor quality signal (traffic data, twitter data, etc) behind them. On the potential hosts side of things, some may sign up for Ohours and start hosting, but since the product is so early and rough around the edges, I doubt they would enjoy the product or have the patience to stick around and host again.
No, in terms of Hosts, the next logical user (or users, as things turned out) for Ohours were David Tisch, Christina Cacioppo, and Ben Siscovick.
Why? Mostly because we are all young, accessible startup mentors who know (or know of) each other, are pleased to take part in the same movement (boosting NY tech primarily, seeing Ohours do well secondarily). Most importantly -- while Fred I imagine most certainly wants to see me succeed with Ohours, Vinny, David, Christina and Ben all have the time and care (for one of their own!) to give me endless feedback on how to make the product better -- better for them... specifically!
By the end of Day 1 (more or less) it was clear that those folks were my core. All the next features I was building for them and only them, and since then, that core group has expanded very logically from Jr VCs to Jr VC-meets-enterpreneur in Jordan Cooper, to "experienced" hacker/founders like Mike Horn.
When you look at the network of Nate, Vinny, Christina and Ben, well Jordan and Mike are pretty logical, though different next nodes. Fred Wilson, no matter how great a guy, is not yet a Next Logical Node, and would have been a major distraction.
Twitter found this lesson out the hard way.
The Oprah Effect
Back at the Web 2.0 Expo in NY, in a session called Turning Users into Active Users, Twitter's Josh Elman talked about "user accounting" at Twitter and how they measure the health of a specific user and the health of the network (an earlier version of the deck on the same topic is on SlideShare).
One specific example he cited was "The Oprah Effect." As it turned out, while Oprah joining Twitter brought hundreds of thousands of signups in one day, when the team went back to look at the data, they found that nearly none of these users ended up staying around. After that, Josh made it seem, Twitter stopped trying to get press and big new user sets, and instead they seemed to focus more on making sure each new user, no matter where they came from, had the ideal on-boarding process and "converted" into an active user.
What's the new on-boarding process? Of course it centers around creating network density. Twitter saw the obvious correlation between how many people you followed, how strong your affinity was for them, and your chances of becoming an active user, and made changes.
Remember the old "Suggested Users" screen? That got pushed aside for "Find Your Friends."
The lesson from Twitter is a lesson we should all take. Oprah can't help you. In fact, with headaches coming from sudden user spikes and a slew of now dissatisfied users, you'd be hard-pressed to find a reason why the Oprah Effect was actually good for the company.
I imagine this lesson is broader than just what happens when Oprah talks about your service. Any attention given to adding additional nodes which are not the most logical next nodes ultimately lowers network quality and hurts your chances of getting those millions of active users you always wanted.
Facebook, Twitter, Foursquare... For the most part these services grew from very small, passionate groups of users which could not be described as anything but the right fit for the network at the right time.
It didn't make sense for Dennis and Naveen to have people from San Franciso -- no matter how influential -- in the first batch of users just like it didn't make sense for students from non Ivy, elite, or Boston-based Universities in the first batch of Facebook schools. And before Ashton Kutcher made PR sense for Twitter (Jan 16, '09), MC Hammer (May 27, '07) was rockin it on the network, which made sense, considering MC is more a tech guy himself, "got" the existing network, which "got" him in return.
If Facebook teaches us anything, non-logical nodes are so "not" worth the attention that it's okay if they join other networks and use similar services while waiting to be the Next Logical Node on yours. Think about MySpace: they blew past Facebook in terms of early users. While MySpace focused on network scale, Facebook focused on network quality, and within short order (short-order was still years!) MySpace users were packing up their digital guitar cases and moving over to Facebook.
So, for the rest of you/us who aspire to build services which millions of end up loving, start with solving the problem of one user -- yourself -- and then solve the problem of another person, a real person, with a name. Then, as you grow, think about the density and intensity you want your network to have when it reaches 100 users, and then 1,000 users, and eventually 1,000,000 users, and then start adding, one by one, the Next Logical Node.
Lastly, remember this: if, through this process, you find that the product stops working for you, or the second person, and the next logical person, just stop. It's always the right decision to stop, get it right -- make something 1, then 5, then 10 people can't live without -- and then only bring in the next node when you've built something they too can truly love.