Taking off: How SumOfUs and CrowdTangle grew to scale
In the categories:
- For Startups
How Crowdtangle sold to Facebook and SumOfUs.org acquired tens of millions of users
You’ve founded a startup. You’re starting at nothing and trying to get “to scale” — whatever that may mean in your part of the world. So, how are you going to get there?
First, I’m going to talk about growth a bit in the abstract, through some very useful frames provided by The Lean Startup (which you should read if you haven’t already — start with the “Grow” chapter!). Then I’m going to give two case studies of growth from New Media Ventures portfolio members — a first-person account of my experience at the organization I founded, SumOfUs.org, and an interview with Brandon Silverman of CrowdTangle.
What you measure matters
The first thing to know about growth is that it’s incredibly important to choose the underlying metric carefully. There’s a famous management adage: “You can’t manage what you don’t measure,” and it applies in spades to thinking about growth.
For instance, if you are a digital campaigning organization, and the metric you focus on growing is number of email list subscribers, you will think about growth quite differently than if you choose an underlying metric of number of people who have taken some action during the last month, vs number of people who have watched one of your videos on social media, vs number of people who have signed up to make a monthly donation. If you’re a SaaS (software-as-a-service) company, the underlying metric might be number of unique customers using your service daily, number of paying customers per month, or total monthly revenue.
Any of these could be the right metric to focus on at a given point in time in your organization’s development, and if you get big enough you’ll probably want to track and assign responsibility to different staff/teams for more than one of these. But early on as an organization you should pick just one and really focus on optimizing it for months or even years at a time.
Your engine of growth
Once you’ve picked the metric you want to focus on growing, then you should think carefully about your “engine of growth.” The Lean Startup defines an engine of growth as “the mechanism that startups use to achieve sustainable growth” — where “sustainable” means that one-time activities (like a big round of advertising paid for by a large incoming investment or grant) don’t count. There are three possible engines of growth:
The sticky engine of growth. This is when most of your growth comes from repeat customers. So, for instance, if you’re measuring number of hours spent on your news website, it may be that most of those hours are from the same people coming back over and over again — and your biggest potential engine of growth is via those same people coming back more often and for longer.
The viral engine of growth. This is when most of your growth comes through new customers who, one way or another, heard about your product/service from existing customers. For instance, let’s imagine you’re creating a secure messaging system for activists to use during marches and rallies. Once I have downloaded your product I have an incentive to make sure my friends also download it, so that I can use it to communicate with them at the next march or rally. Similarly, advocacy organizations optimizing for email subscribers generally grow through existing subscribers sharing petitions with other people, who then sign those petitions and share with their friends, etc.
The paid engine of growth. This is when you have to spend money per unit of growth — but that unit of growth also yields revenue that can be used to grow further. For instance, if your product is a CRM (constituent relationship management software) that nonprofits purchase via monthly subscription, you might have a sales team that averages $1000 in cost per new customer, but each new customer brings $3000 in revenue over the first year that you can plow back into the sales team, thus allowing you to bring in 3 more new customers — and so on. Or if you’re a text-message-based campaigning organization, you might buy mobile ads targeting new potential grassroots activists to sign up for your SMS list. If you can acquire a new SMS subscriber for just $1, and then get each of them to donate on average $3 in their first month, you’re going to have a great year!
Just as you should focus on one metric at a time, early on you should usually pick one engine of growth to optimize at a time. In other words, as a small startup, don’t try to optimize your word-of-mouth virality and your paid advertising strategy at the same time. Pick whichever feels more promising, or like lower-hanging fruit to test, and throw yourselves into it until it feels like you’ve exhausted its potential.
The SumOfUs growth story
SumOfUs is a digital campaigning organization coming out of the MoveOn mold, which I founded in 2011. New Media Ventures was one of our earliest funders, and in total we raised a funding round of $175,000 through NMV’s own fund and its network of progressive investors. Below is a narrative of our growth path. You can also find a presentation covering some of the same material, with more charts, here.
We started out with a few list seeds I negotiated that totaled something like 250,000 email list subscribers. We had a viral engine of growth — primarily via our existing subscribers sharing our petitions on Facebook, and their friends then signing those petitions.
Over the following 1.5 years, we grew our active English-speaking email list at a gradual underlying rate, with a few spiky growth moments (each of which only lasted at most a week or so, stemming from a single petition) punctuating the curve. At this point we had a bit over a million subscribers.
We then hit on a big hockey stick growth curve for about 5 months. Basically, we discovered that something had changed in Edgerank — Facebook’s algorithm for promoting content in users’ newsfeeds — and we were able to leverage that change into a vastly higher viral coefficient. We didn’t know how long the change would stay in effect, so for a few months we shifted staff time and resources into finding content that took advantage of the change, which in retrospect was an extraordinarily smart decision. We called this period our “growth surge.” By the end of it we had approximately 4 million subscribers — i.e., we had grown more than 3-fold during that 5-month period.
Our English-speaking list growth then stalled (for reasons that still aren’t fully clear, though further changes to Edgerank were certainly a contributing factor). In the new environment, where we had a lot of new English-language users but couldn’t easily acquire more, focusing on the sticky engine of growth felt very important. Shortly after it became apparent that the conditions for us to grow our number of subscribers quickly had changed, we decided to switch away from subscribed users to a new metric: monthly Members Returning for Action, or monthly MeRA. This was defined as “the number of unique users per month who take an action that is not their first-ever action.”
At the same time, we started putting resources into our viral engine of growth towards a new audience: German and French speakers. Now, this violates my advice above about focusing on exactly one engine at a time. But at this point we had about 30 staff — and when you get large enough, you can and will probably have to both walk and chew gum at the same time. The sticky engine was keeping our active English-language user base high, but the vast majority of new users were coming in from these other languages.
The key factors to our success in growing quickly were:
The key metrics I wanted to optimize for (new users per week/per petition) were embedded in-your-face throughout most staff’s daily workflows. So people were getting constant feedback on whether what they were doing was contributing to the cause. We were able to identify very quickly when underlying conditions (Edgerank) had changed in our favor because these metrics were so visible.
We invested early in technology that made it easy to run optimization tests on the key component of our growth engine (Facebook share text and images). During the period of our “growth surge,” lots of media outlets were reporting much higher virality coefficients via Facebook shares, but other digital campaigning groups weren’t. I think the main difference between us and the other groups was that Facebook share text testing was already embedded as routine practice throughout the organization. This had already paid off before the changes to Edgerank — it magnified our growth across our viral moments — and set us up for massive growth when Edgerank changed.
We were humble about our ability to predict what would go viral, and set up very lightweight approval processes that allowed us to throw a huge amount of petition spaghetti at the wall. It felt almost inevitable that some of those petitions would stick/go viral — and in practice this included many petitions that we would not have predicted in advance.
When we started focusing on multiple engines of growth at once — the sticky engine in English, and the viral engine in German and French — we put up internal firewalls between the two engines. I.e., the team focused on viral growth in German and French only rarely had to coordinate their campaign plans with those who were optimizing for MeRA to our English-language list. Each group had its own single, clear North Star metric.
The CrowdTangle growth story
CrowdTangle was a for-profit SaaS company, also founded in 2011. New Media Ventures invested in CrowdTangle in the same 2011–2012 cohort as SumOfUs, and Facebook acquired CrowdTangle in 2016. I interviewed Brandon Silverman, founder and CEO, to glean lessons from how he grew the company.
Q: What did your growth trajectory look like?
A: When we were bought, we had been doubling our revenue every year since we launched. And we were seeing double-digit MRR (Monthly Returning Revenue) growth every quarter at least, and often every month. Our average contract size was growing, our number of clients was growing. We had a good product that was unique in the market, and we had a good reputation. It was product-market fit.
Q: What was your primary growth engine for acquiring new customers?
It was all word-of-mouth. We grew almost entirely through inbound inquiries — the social media manager at the New York Times is friends with the social media manager at Slate, etc. That’s especially true in a tight and insular space like the New York City media sector.
Meanwhile, Facebook was also recommending us to folks, which helped. We were solving a lot of pain points for people they cared about, and doing it fairly cheaply. They would meet with a publisher who would say, our #1 request is X, and Facebook would say, well Crowdtangle will do that for you.
If word of mouth is your growth engine, there are network effects. More people using it = more people recommending it. And there was plenty more room in the market still when we were acquired. When we pitched investors, we described our market opportunity as every active page on FB — which we were nowhere near having tapped out.
Q: What were three things you did to foster growth?
A: First, we priced for customer growth, not for financial growth. We left money on the table in terms of how much we could earn from individual contracts, in order to get market share in the verticals we were targeting. Our space was trending toward a commodity-based space — it’s hard to differentiate yourself when you’re just using public data. Initially we tried to attract large, high-profile clients with the goal of building momentum, capturing important parts of the market and hoping we could continue to solve more and more problems for clients over time.
Second, focus in what we provided: We never did any consulting, any service-based work. All you could buy from us was a seat in the dashboard. That let us focus on the product itself.
Third, in terms of financial growth: The product itself and the pricing model meant that larger companies had larger licenses, and as we grew we started to land more and more six-figure contracts. We didn’t charge per user, we charged per team — a team could have an unlimited number of seats. But larger companies needed to purchase licenses for multiple teams, because the product wasn’t designed for large numbers of people on the same team. Our contracts averaged about $12K/year, but by the end we were signing a number of six-figure deals a year. We were moving towards larger licenses, larger packages for larger companies with more teams.
Q: What hard decisions did you have to make?
A: Pricing is really hard. I in no way thought we had figured out pricing at the end. I don’t think you ever figure it out, you just make sure you are always talking about it. We had a larger theory about it being a competitive space, so we underpriced in order to get market share.
When we were bought, we were about to come to a critical moment as a company. At any SaaS company, you have to decide if you’re going to go upstream and sign your existing clients up to larger contracts, or expand to more clients. We were literally about to face that. Once you start signing 6-figure deals, you have to lean into that heavily, at which point you have to sign a sales team that’s good at that stuff… Or you have to back off of it and just go for breadth.
This article is one of a series of pieces I’m publishing with New Media Ventures, to support their investments in a new cohort of 17 startups doing vital work building progressive political infrastructure and championing progressive values.