Blog

The Friend Multiplier

The second in an occasional series about product design heuristics. The first article was about mutual benefits.

Another heuristic that's useful in product design is something I call "The Friend Multiplier". There's two parts to it, and both are requirements:

  1. Your product must be useful on its own.
  2. Your product must be n times more useful if a user's friends use it too.

It has a powerful effect when your product is inordinately more useful (or more compelling, fun, deep, etc) when a user's friends also use it. Done right, it feels like an entirely new dimension opening up: a new world of possibilities, not a couple of new features1. It's important that the benefits are genuine (there's nothing magical about an artificial barrier of "Wait 4 hours, or invite 5 friends to continue"), and that the friends are a user's real friends2 – people who would have a coffee with your user, given the chance.

But don't overlook the first part: few products can get away with being useless for solo users. Users are more likely to stick around if it's easy for them to start using your product3, and everybody wants to try something before recommending it to their friends. And what if they're not your ideal user? What if they don't live in a big city, or lack early-adopter friends? What if they've got a 4 year old phone or a 10 year old laptop? Will your product still work well for them? Millions of people fall into this demographic and they're often neglected, especially by tech startups. They might not be your primary audience, but if you improve their experience then your product will be better for everybody.

But the multiplier is the magic part; the reason this heuristic indicates your product is on-track, and that word-of-mouth growth is plausible.

Some examples of friend multiplier magic

I really like FourSquare as an example of the friend multiplier, despite their recent disoriented product decisions. You use FourSquare to log where you've been4 and it recommends new places to you. That's mildly useful & fun on its own, but the day FourSquare really clicked for me is burnt into my brain. I was in a café that I'd never visited before and I checked in on FourSquare. FourSquare popped up a 2-year-old note from a friend, giving me the wifi password. Magical! My friend hadn't left it for me; it was a generic tip. But FourSquare knew we were friends, and knew that my friend had left a note here, so it showed me the note when I checked in. There's no wizardry in the technology powering it, but the effect was profound. I was now aware of hidden traces left by my friends; traces I could stumble upon, seek out, or leave for others.

LiveJournal is a simpler example. LiveJournal users post diary entries to the web, either publicly or privately. Which is useful in itself, if you want to keep a diary. But when your friends are on LiveJournal too, you get an intimate window into their lives. LJ's fine-grained privacy settings mean your friends have control over what they're sharing – but you'll know more about how they feel, what's on their mind, and what they're doing. I never had a LiveJournal, but a lot of my university friends did. I always had a sense that there was an in circle – a deeper, mutually supportive, more emotionally bonded group – that I was outside of. Again, there was no groundbreaking technology involved. But the social impact on individuals was huge.

It won't last forever

It's important to have a target number in mind for your ideal user's friends. The magical feeling normally ebbs away after a (normally low) threshold. Twitter's a rare example with a high threshold: it's better when you follow a few hundred people. Most products have more in common with a shared calendar; more than a handful of people using one results in a deluge of events, making it hard to find ones relevant to you & making the calendar feel like a mess. So pick a low number & design with that number in mind.

Some counterexamples

This heuristic isn't a law. Lots of products thrive without any kind of social features, and (as before) there's lots of products where it's best avoided. Sometimes that's immediately apparent: I don't want my friends involved in my banking or my to-do list. I think that it's best to keep them away from dating & relationships, too; Tinder's "friends you have in common" feature always felt like a warning about potential future awkwardness, rather than an endorsement.

There's also fuzzier cases. Features that seem like a good idea can fall apart like wet cake once the complications of real life get involved. Windows 10 has a feature where your friends can log on automatically to your wifi. Which sounds great – your computer already knows who your friends are, so why not let them onto your wifi without having to ask for the password? Ah, but the details: your computer doesn't know about your friends. It knows about your contacts. Contacts contain friends, family, colleagues, ex-lovers, stalkers, businesses, abusers, toxic people you've cut out of your life. A whole lot of people you'd want kept away from your wifi network.


The friend multiplier is a simple but effective guideline. It's not a requirement – if it doesn't make sense for your product, don't worry about it. But in an increasingly connected world, it makes sense to think about how that can make your products much more compelling.


  1. A "solo" vs "with friends" feature comparison would be missing the point, anyway: product design is about what your product lets users accomplish, not the bullet points you can list.  ↩

  2. "Real" isn't the same as "real life"; internet-only friends count for this too. It's about genuine, two-way friendships – not people followed because a user likes their jokes, or stuff they make.  ↩

  3. Or service – especially services. If it's easy for someone to start using your service it's more likely they'll find value in it, so it's less likely they'll cancel. Fewer cancellations mean a lower churn rate, and a lower churn rate makes it easier to build a viable business.  ↩

  4. Which is interesting in itself; "Wow, have I really not been here for a year and a half? Really?"  ↩