Monday, September 22, 2008

The three drivers of growth for your business model. Choose one.

Master of 500 Hats: Startup Metrics for Pirates (SeedCamp 2008, London)

This presentation should be required reading for anyone creating a startup with an online service component. The AARRR model (hence pirates, get it?) is an elegant way to model any service-oriented business:
  1. Acquisition
  2. Activation
  3. Retention
  4. Referral
  5. Revenue
We used a very similar scheme at IMVU, although we weren't lucky enough to have started with this framework, and so had to derive a lot of it ourselves via trial and error. Dave's done a great job of articulating the key metrics you want to look at in each of these five areas, and I won't bother repeating them here (go read the presentation already). He also has a discussion of how your choice of business model determines which of these metric areas you want to focus on. That's where I'd like to pick up the discussion.

I think the salient question to ask about any business model is: what is the primary driver of growth? I break the answer to that question down into three engines:
  1. Viral - this is the business model identified in the presentation as "Get Users." Here, the key metrics are Acquisition and Referral, combined into the now-famous viral coefficient. If the coefficient is > 1.0, you generally have a viral hit on your hands. You get increasing growth by optimizing the viral loop, and you get revenue as a side-effect, assuming you have even the most anemic monetization scheme baked into your product. The law of large numbers (of customers) says you can't help but make at least some money - your valuation is determined by how well you monetize the tidal wave of growth. Examples of this are well-known, and (in my definition) include any product that causes new customers to sign up as a necessary side-effect of existing customers' normal usage: Facebook, Myspace, AIM/ICQ, Hotmail, Paypal.

  2. Paid - if your product monetizes customers better than your competitors, you have the opportunity to use your lifetime value advantage to drive growth. In this model, you take some fraction of the lifetime value of each customer and plow that back into paid acquisition through SEM, banner ads, PR, affiliates, etc. The spread between your LTV and blended CPA determines either your profitability or your rate of growth, and a high valuation depends on balancing these two factors. To the extent that you have good word-of-mouth, activation or retention, these factors tend to drive down your CPA or drive up your LTV, and so are nice bonuses. But because paid traffic is fundamentally a bidding war, it's important that you have a differentiated ability to monetize customers better than other people who are bidding for the same traffic. Otherwise, your CPA will get driven up close to or exceeding your LTV, and you can't grow profitably anymore. IMVU is in this business, as is Amazon, Netflix, Match.com, and CafePress.

  3. Sticky - Dave calls this "Drive Usage" and I think it's the model that causes the greatest confusion. Because Activation is a key term in the Viral business model equation, and Retention is a key term in the Paid business model, it's easy to mix up this type of business with the other two. For example, you often hear eBay or Neopets described as having viral growth, but I don't think that's correct. What those sites have in common (despite their very different audiences) is that something is causing their customers to become addicted to their product, and so no matter how they acquire a new customer, they tend to keep them. This has led to exponential growth. For eBay, this is caused by the incredible network effects of their business (so-called demand-side increasing returns and supply-side increasing returns). For Neopets, it's simply a side-effect of their game-like product design. Either way, you can use any marketing channel that's available to bring in new customers, including word of traditional advertising, SEO, SEM - wherever you can find prospects who are going to find your product addicting. But it's not really viral growth, even when it's exponential. Again, looking at eBay - most buyers and sellers would be 100% happy with eBay if it already had a critical mass of people to bid or create auctions. Although many eBay fans love to tell their friends about it, they really don't have a need to bring them on board. As far as they are concerned, that's eBay's job. That's why eBay advertises on search engines, and Facebook doesn't.
In my opinion, every startup needs to "pick a major" among these three drivers of growth. It's simply too hard to focus on more than one. It's a choice that has to be made at the level of strategy; notice how similar the tactics are between them. All three probably make attempts at world-of-mouth marketing - it's just that for Viral, it's life-or-death. Similarly, it probably makes sense for everyone to take advantage of SEO (hey, it's nearly-free traffic). But a Viral company who is focused there is probably going out of business.

The difficulty is exacerbated by the fact that these models also cut across business functions. Sometimes we have the attitude that the Product Development team is the one responsible for Activation and Retention (hey, a great product would do that naturally) or that the Marketing team is responsible for Revenue and Referral (hey, go get me some money or free customers already). In reality, the key metrics for your growth drivers have to be jointly owned. They cannot be delegated, unlike the minor metrics, which can easily be owned by one part of the business, or even outsourced. For example, it's always nice to have someone constantly optimizing your SEM accounts, driving down your CPA. They might even occasionally make "optimizations" that improve CPA but negatively impact LTV (or vice versa). But if you were using the Paid driver of growth, you just outsourced your heart while it was still beating. Oops!

One last thought. Beware the new hire who has "extensive experience" in startups or big companies - using a different growth driver. Make sure you test to see if they are truly open minded, because otherwise you risk them banging their head against a wall, trying to use the tactics that worked so well in their previous company. Be-double-ware of the new hire who has "years of industry experience in multiple companies" all in a different growth driver. It's the rare person who truly understands not just what worked in the past but why.

0 comments: