Startup Metrics for Subscription Web Businesses

One of the most frustrating things about running an online software business that's subscription-based is that it's really hard to tell whether changes we make are having a positive of negative influence on our customer's happiness, and ultimately on our bottom line.

The following are some key metrics that can help you make sense of your business, and if you track these over time, you can see if you're headed in the right direction:

Customer churn (Churn): People who don't renew their subscription.  This is easy to measure month to month, you just count the number of subscribers who don't renew.  To get a percentage, you divide the number of cancellations/non-renewals by the starting count of subscribers.  For example, if you start June with 100 subscribers, and 5 leave over the course of the month, your churn for June would be 5/100 = 5%.   The significance of churn is that you need that percentage to be lower than the growth, or you'll be out of business pretty quickly.


Customer growth (Growth): People who subscribe - woohoo!  This is also easy to measure month to month, you just count the number of new subscribers during the month, and divide by the number you started with.  For example, if you start June with 100 subscribers, and 10 new ones join over the course of the month, your growth for June would be 10/100 = 10%.  This number should be higher than churn.  It's also one reason the expression "If you're not growing, you're dying" isn't just an abstract concept.  In a subscription business, it's the hard truth.

Customer Acquisition Cost (CAC): How much it costs to acquire a new subscriber, including advertising, followup, discounts, etc.  CAC for a given month would simply be your total marketing and sales costs across all channels, divided by the number of new customers acquired that month.  So again, if you spent $2000 in June, and landed those 10 new customers, your CAC is $200 per customer.  Is that good or bad?  See the next two items below to find out.

Average Revenue Per Customer (ARPC): How much you earn per customer in a given time period.  This one is easy - we figure out the total revenue from customers in a given month, then divide by the number of customers.  If we make $1000 in June from our 100 customers, then our ARPC is $1000/10 = $100.  More on this number below.


Lifetime value: How much an average customer will spend with us over the course of our relationship. This is a harder number to come by, but it's based on the Churn number from above.  Lifetime value (LTV) is useful to understand, because you can use it to decide how much to spend to acquire a new customer (hint: the cost to acquire should be LESS than the LTV of a customer).  If you're in the early stages of a startup, though, you really don't know how long your customers are going to stick around yet.  You probably need at least a year's worth of data before you have a good sense.  But that won't stop us from trying - here's a quick and dirty formula for estimating lifetime value: 

LTV = (1/Churn) * ARPC

 
So what this means is that we can get a decent guess at LTV based on our churn, and average revenue per customer.  This will shift over time, but it's better than nothing, and can be surprisingly accurate.

In our example, if we've been operating for 6 months, and have 100 subscribers, and the average monthly churn is about 5% (in June we lost 5 of 100), and we're making about $100 per subscriber monthly, our LTV looks like this:

LTV = (1 / 0.05) * $100 = $2000

That's pretty good!  With that number ($2000 LTV), we might be able to spend a little more to acquire a customer.  To leave a large margin for error, we might decide to spend up to $500 to acquire a customer.  Of course, we'd first have to make sure any marketing channel we chose will actually deliver the right kinds of customers, and that they will convert at the rate we expect.  Uh oh, we haven't talked about that yet...

Conversion rate:  What percentage of our new leads will become customers?  If you offer trial versions, then you'll have two other values to consider - the lead to trial conversion rate, and the trial to customer conversion rate.  Knowing your conversion rate, not just overall, but also by source (referrals, Google ads, magazine ads, search engine results for specific keywords, etc.) can give you insights into where to focus your marketing resources. 

You can use all of the above to project new subscribers, churn, and revenue for your business, updating it with actual values on a monthly basis to keep it accurate.  Here's a sample set of projections you could make (there are others, but hopefully this demonstrates what you can do with these metrics):

Est. New Subs = Growth * Existing Subs

Est. New Revenue = Est. New Subs * ARPC

Est. Retained Subs = Existing Subs - (Churn * Existing Subs)

Est. Recurring Revenue = Est. Retained Subs * ARPC

These numbers are a good start, but they are missing something - a correlation to the state of the product or service.  For example, in the very beginning of a startup product or service, you might expect churn to be a little higher, simply because the product isn't that great initially.  But I'm out of room, so that will have to wait for a later post.

How are you measuring business performance?

Product Planning with Remote Stakeholders

One of the things I've struggled with over the years is how to get people really engaged in the product planning process.

For software products, this usually means having them prioritize features, or possibly group things into releases or timeframes.

This is hard enough if you've got everyone is the same room, maybe scribbling on a whiteboard, or pushing around sticky notes, but get that much harder when people with important input are in another location, or just too busy to pin down.  In our case, we're a virtual team, so we're almost always in different locations as a matter of course.

To solve this problem, we've tried typical project management tools, spreadsheets, online docs, and other possible solutions, but they've been either too awkward to use, especially remotely, or worse, they don't let you rank things naturally, so we wound up with lots of "top priority" features.  As in "Everything has to be in the first release."  Not helpful.

So years later, I've still got this problem, and it really started to annoy me.  But wait, I've got a software company! So we put something together that would solve our little problem by making it really, really easy for people to provide useful input, and not wind up with an unranked list of things.  Even better, we could actually all participate from wherever we were, with just a web browser, no special software to install.

The idea: an online collaborative "board" split up into columns, with each column representing a release.  We then add feature "cards" to the columns, like "New pricing widget", and can then drag and drop the cards to order them, or move them to different columns to plan the release. 

The big win here though is the real-time interaction - everyone can use a standard web browser to use the board, and see updates within a second or two.  We all get on a conference call, then join the board so we can see what everyone is adding or moving in real time.  It's really simple, useful, and we get much better engagement from everyone than the old way (circulating a spreadsheet to rank things).

There's also some other cool variations on this like tracking task status for a release by using the columns as "To Do", "In Process", "Completed" instead of release names, etc.

So now that I'm pretty happy with that, I couldn't help wondering if anyone else might benefit from this concept.  So we're starting a private beta to see if other product managers or teams that aren't all in the same place could use this to plan and prioritize things more effectively.  If you're interested, you can signup here.

Startup Lessons Learned Conference

Just got back from the wonderful Startup Lessons Learned conference in San Francisco, put on by Eric Ries.

This was the first major gathering of people interested in the new "lean startup" movement, which basically ties together the ideas of Steve Blank on Customer Development with ideas from the lean software/Agile development community.

Although the educational information presented was great, I had already internalized a great deal of it from my own readings and experience. 

But the case studies were really the heart of the conference.  Presentations from Dropbox, PBWorks, and a variety of other startups in different stages really illustrated the principles much better than any formal theory could.

In particular, most of the folks presenting case studies had to "pivot" their business model when some aspect of their original vision just didn't work.  Some did this more than a few times before getting it right (or close to right). I was struck that almost no one reached critical mass through advertising or PR.  Instead they started with a "minimum viable product" and attracted referrals as they went, sometimes from unexpected sources.

For me, the biggest takeaway was a renewed commitment to connect deeply with customers, better understand the problems and goals they have, and to do the best we can to solve them as simply and effectively as possible.

Oh, and it made me want to blog again :-)