[mastermojo]

The 5% churn = 79% annual churn is incorrect. Churn doesn't compound, because after churn your user base is smaller and 5% of that becomes a smaller number.

It is actually 0.95^12=0.54 (5% churn/month = 46% annual churn).

[one-more-minute]

That formula is more correct, but only assuming you don't replace any churned customers – or at least, replace them all at once at the end of the year, which seems like a pretty unreasonable assumption to me.

If you replace customers every month then the number leaving per month is constant, so you end up churning 0.05×12 = 60% of your userbase per year.

Edit: The 12x model is also pretty coarse, but you can take the limit of continuous replacement, and you end up churning

    -12log(1-x) = 0.616

× your customer base per year.

[meesterdude]

ya'll are total nerds. Have some upvotes!

edit: subtle sexism

[cperciva]

Churn doesn't compound, because after churn your user base is smaller and 5% of that becomes a smaller number.

There's a more fundamental flaw: Not all users are equally likely to be lost. If 10% of your users have a 50% chance of finding mates and leaving the dating site each month, while 90% of your users are hopeless losers who will never find love, you'll have a 5% monthly churn rate, but only a 9.9998% annual churn rate.

[shalmanese]

The even more corrosive case in dating is that there's no "happy path". For a SAAS product, you might have incredibly churn for users who join just to kick the tires and leave quickly afterwards but you can sustain yourself on a core of extremely long terms customers who will almost never leave and keep paying reliably month after month.

With dating, there's no equivalent case. Even hookup apps, your users eventually get tired of the hookup lifestyle or they age out or they get enough casual partners to satisfy themselves.

Dating is one of the few services where your customers get unhappier the more money they spend on you.

[abandonliberty]

You're assuming no replacement.

I think the calculation runs like this as per[0] 5% => 1/.05 = 20 months average lifespan = 1.67 years => 1/1.67 = 60% yearly churn.

[0]http://en.wikipedia.org/wiki/Churn_rate#Churn_Rate_of_a_Cust...

[andrew_null]

Thanks. I did totally botch the math, I typed this up in one go and mixed up the formula, thanks for pointing it out. Updated to be more accurate.

[teej]

Realistically you are adding more users each month. A business that is not adding new users each month is going to be out of business.