I’ve been thinking about how distributed applications will grow. This post will cover why referrals are powerful, how to get more referrals, and how crypto tokens create a double network effect that can supercharge referral results.
Why referrals are powerful
I worked extensively on referrals as a founder of Reveal Chat (B2C, grew to millions of users via referrals and was acquired by Napster) and at Gusto (B2B, growing fast and on track to become huge).
Like any growth tactic, referrals aren’t a panacea, and it can be tempting to prioritize them too early or too aggressively — I’ve done both. Make something people want. Don’t pour water in a leaky bucket. Never spam. All true. But when a well-executed referral program is working, it can be transformative and exhilarating. Like this:
Hockey sticks aside, I love referrals for a few reasons.
- Referral growth is a proxy for product quality. It’s not equally true across the board. Some highly referred products are just optimized spam cannons, and some get better for everyone when new users join (network effect!), while others don’t. But there’s a correlation between how much users like your product and how willing they are to share it. Waze CEO Noam Bardin says a growth team’s job is to “accelerate organic trends.” And word of mouth is the best organic growth channel in history.
- Referred users are the best kind. They’re more likely to sign up, enjoy your product, pay, stick around, and refer others — it’s a virtuous cycle. And since paid acquisition is generally efficient, which is a nice way of saying you need a massive budget to drive massive growth, referrals can slash your customer acquisition cost (CAC). This is especially true if you aren’t offering users a direct incentive for referrals (Facebook), or if you’re offering something that costs you very little (storage for Dropbox) or requires a purchase to redeem (discounts for Airbnb).
- Referrals are fun to think about. I’ve always tested 50/50 right brain-left brain. And referral growth is a math problem — that you solve with creativity! How cool is that? Let me show you what I mean.
How to Get More Referrals
I made a simple framework called PICs to help me think about referrals as a math problem - because nothing spreads faster than PICs.
Percent of users who invite or share *
Invites or shares per inviter *
Conversion rate on invites or shares
= viral coefficient AKA K factor AKA the number of additional users referred in by any single user.
- If you can sustain a viral coefficient above 1 for some period of time, you will grow exponentially and may end up with one of those crazy stories where you took off on a plane from SFO and landed at JFK with 500x more users, and hopefully zero crashed servers. This is what most people mean by “going viral.”
- Your viral coefficient is the product of P, I, and C, so increasing any one of those numbers by 20% will increase your viral coefficient by 20%. So the question you might ask is “which of these numbers can we increase by 20% with the least work?” Incidentally, there are natural ceilings on P and C. Unless sharing is required to use your product (e.g. Snap) it’s typically hard to get more than 10% of users inviting or sharing. It’s typically hard to get more than 25% of people who receive a private invite to become users, or to get more than 0.25 users from your average public share. I is the most “uncapped” number — it’s often easier to increase the average invites or shares per inviter than the other two numbers. This is why address book import and “invite all” functionality were prioritized by Facebook and LinkedIn to power some of the most aggressively performant referral growth flows in history. Haven’t signed up for a new account in a while? These flows are still the centerpiece of onboarding for both apps.
- One channel for invites or sharing is likely to work better for you than others — probably much better. What do you want your users to do? Invite friends privately via email/SMS? Share publicly via Facebook/Twitter? Post to online forums? You can and should run small scale tests to find out what converts.
So that’s the math. The numbers in B2C and B2B and across products in each are very different, but the PICS framework holds. That’s the great thing about math — it’s the same everywhere.
The creativity is in figuring out how to drive up P, I, and C without compromising your users’ experience. And that’s about understanding your product and your users. Here are some questions to ask yourself, or better yet, to talk through with your users and your team.
P: Why would users want their friends to join your app? Is it more like a party or a hardware store? People are more likely to refer friends to your app if it makes them look cool or smart, or if everyone involved will benefit from enjoying it together, like a party. If your app doesn’t get better for users when their friends join, it’s more like a hardware store — some might tell friends about it if they love it, but increasing the percent who do will be tough, and you may need to consider tangible incentives — like a cash payout when someone they’ve referred makes a purchase.
I: Who do you want your users to refer? How can you make that easy? Relevancy matters: the number of people I know who might be interested in a new social network is larger than the number of people I know who might need a payroll solution. At Gusto, we learned from our users that many of them wanted to refer more friends, but they weren’t sure which of the people they knew would be great customers for Gusto. So we asked them to connect their contacts and analyzed the email domains to recommend the best people for them to refer.
C: Does the referral copy clearly explain your value prop? (Hint: find out how users tell friends about your product in person). Are you making it as easy as possible for referred people to sign up and start getting value? Many mobile apps grow via SMS, which has fast and high open and click rates but relatively low conversion rates because of all the friction between a click and an app install. Web apps often enjoy lower signup friction and thus a higher C.
How crypto tokens create a double network effect that can supercharge referral results
Okay, here’s the fun part. Think of your favorite app that has a traditional network effect: your experience gets better when more people are using it. Let’s use Medium as an example. When new users sign up and start posting, clapping, and commenting, the network gets better for you personally — especially if they’re your friends. That might make you want to tell the writers you know to join Medium. Network effect #1 at work.
What if Medium started giving you limited-release “MediumCoin” just for being an awesome user? What if you could buy or sell MediumCoin instantly for cash, or spend it on Medium to get access to special features, or hold it with confidence that its value would increase if the network continued to grow? What if you and your friends could get MediumCoin free if you referred them and they signed up? With your personal and economic incentives aligned, now you might REALLY want to tell your friends to join Medium. You might even start to sound a little crazy when you talked about it. That’s network effect #2.
People are already testing the potential of coin-powered referrals. ICO marketers offer token bounties to users who complete simple marketing tasks (e.g. post a link on Facebook, follow a profile on Twitter, loan out your avatar or signature line on Bitcointalk.) These are proto-referral or affiliate marketing programs that can be evaluated and improved using frameworks like PICs. They may only be scratching the surface of the power of this approach.
In the early 2000s, PayPal achieved 7 to 10% daily growth via referrals by giving both people involved $20 when a new referred user signed up. But a key limitation of PayPal’s model is that everyone knew what $20 was worth to them. Some people were willing to spend a few minutes signing up to get it, others were not.
When tokens can be traded instantly for cash, the current cash value of a token arguably represents its floor. It can also be spent within applications, or held for potential future value, creating a fear of missing out that may make signup even more enticing. After all, $20 of Bitcoin from 2009 would be worth over $25 million dollars today.
I’ll be testing referrals concepts in new crypto projects soon, and I’m looking for thought partners to help make me smarter. If you’re interested or want to help, please reach out privately or in the comments. I’d love to hear from you.