I posted this answer to Quora and figured I would share it here as well. You can find me on Quora here.
What is considered a significant number of users for a free consumer internet product?
If it’s a mass market product and you are looking to build a venture-scale startup, you need 10s of millions of users, maybe more.
Looking at the end state
To pick an arbitrary end state, let’s say you want to end up at $100M revenue runrate. If this seems to high or low to you in defining a “significant” number of users, then just pick your own number and apply the reasoning below.
So starting with the $100M number, this is why you need 10s of millions of users, typically:
Ad-based business models
If you go with advertising-based models, CPMs are traditionally quite low for mass market products- usually <$1 per 1000 ad impressions[0]. For social sites that number is more like $0.25 CPM.
So if you want to make, let’s say $100M a year, then it’ll take you 100B impressions per year, or 8.3B per month, to build that kind of business. You need a LOT of users- certainly in the 10s of millions of uniques per month who are quite engaged, in order to make that work. You would have a top 50 website to make this happen. We’ll read about you in the news.
You can cut this number down if you manage to create, say, a viable search engine. Then you might have CPMs more like $50-100, which cuts down your ad impression requirements significantly, but then you’re competing with Google. Similarly, you’d need millions of users on your email list to compete with Groupon, but you don’t need 100M email subs to get to a good revenue number.
Facebook has 570B+ pageviews/month[4], which is 5X more than Google, but their revenues are still 1/10 or 1/20 that of Google’s[5][6].
Social gaming business models
The same is true for mass market consumer internet models based on transactions. You end up with about 3% of users converting, and their ARPPU is in the single digit $ figures. So you still end up needing 10s of millions to hit a big revenue number like that.
To generate $100M runrate, you need $8.3M revenue per month. At 3% conversion and $5 ARPPU, that’s still 55M uniques per month.
The way you build a Zynga is you build a company with 266M MAU[3].
Transactional and vertical markets
If you are building a more transactional product, then the numbers above can be significantly increased. For example, if you’re a free consumer internet site for job hunters, then you’re getting a % of a transaction that’s $50k-$100k, so that’s much better.
The downside is that vertical applications tend to have a tough time acquiring and holding onto their users, whereas horizontal sites focused on communication or content publishing usually are viral and hold on to their users. If you have a tough time acquiring or holding onto users, then you eventually pay your margin out to Google, Facebook, etc. and your profits go to zero. It’s a tradeoff.
(Thus the focus of so many companies to take a transactional thing and make it social, to try to capture the social benefits- like social shopping, social job hunting products, etc.)
Just starting out?
So after reading all of the above, you might want to know how likely you are to get into a trajectory to a substantial user number. If you’re just starting out, I might look at the following:
- size of market (do I think 50M uniques/month want to do this?)
- how fast is it growing (could you approach 5k-50k new users per day?)
- have you proven your product out with a sizeable base? (50k-500k active users per month?)
- Are you part of a large existing category of products that has 100M+ uniques per month?
I think the above could all be clues to evaluating your particular product. But you never know :-)
[0] Here’s a more detailed CPM breakdown: http://andrewchenblog.com/2008/0…
[1] Some ARPU numbers: http://giffconstable.com/2009/07…
[2] Some ARPPU numbers: http://andrewchenblog.com/2009/1…
[3] Zynga stats: http://www.appdata.com/devs/10-z…
[4] http://www.businessinsider.com/h…
[5] http://investor.google.com/finan…
[6] http://mashable.com/2011/01/17/f…
