Analyzing churn rates, free trials, and other metrics | Hacker News Reader

Analyzing churn rates, free trials, and other metrics

jlongster.com

149 pointsshinryuu16 comments

> You might think "just crank up the growth until you hit $10k/month". Something weird happens ... you ended up losing money with high growth.

Ratio of new customers showing losses vs old customers showing profits.

Your ratio of new to old customers basically IS your growth rate.

As that ratio increases, you could show "losses". But financially, it's more like investment than losses.

You probably should, in fact, crank up the growth in this situation. Think of it as an investment and calculate your rate of return.

Edit 1: I threw together a quick spreadsheet to illustrate the point (shows cash flows per customer):

https://docs.google.com/spreadsheets/d/19mbsjcwlyLs_KsvPrjdt...

Edit 2: On second thought, rate of return is the wrong measure because time value of money is not significant. What you care about is Customer Acquisition Cost and Customer Lifetime Value.

Edit 3: FYI I'm a developer now, but was actuarial, and Excel whiz w/degrees in finance & accounting. I'm always happy to give anyone a little bit of free help with things like this.

Thanks a lot!

I had a hunch that it was something like that. "loss" isn't really the right word, but I'm focusing on cash flow for the next year. You're right though - it's an investment, but I'd like to figure out the math so I can learn how much cash I need upfront to fuel a specific % of growth.

Checkout your spreadsheet now. don't quite understand it yet, but I'll follow-up with you on twitter (saw your tweets)

Friendly reminder that churn rate isn't a meaningful metric, because churn probability is not constant throughout a customer lifetime: https://medium.com/swlh/youre-all-calculating-churn-rates-wr...

The churn rate will almost certainly go down on its own as your user base gets older, and consist more of long term customers. What you're usually looking for is to compute average customer lifetime, which is better done without assuming constant churn probability, see the linked post.

Don't folks usually track churn amongst cohorts anyways? Churn rate without any associated facts is almost completely meaningless, as you suggest, other than to elicit concern.

Did the OP?

I did not, but I was only focusing on the next year anyway. It's just something to take into account losing customers. If I take a pessimistic churn rate and it goes down over time I'll be better off.

I haven't read the linked post, will do. I don't think I have nearly enough data to know customer lifetime though. I only have 4 months worth of data.

Cannot access spreadsheets now, but limiting free trials to one bank account would be the first thing that I would do. The second is to offer annual plan with a discount. The industry practice is to price annual as 10*monthly price (aka "2 months free"), and play with these numbers. That does introduce another parameter to the model though: the separate conversion rates to annual and monthly (or - how these 6% would be split).

Totally agree with your first two points. I would also suggest experimenting with trial durations. I don't know how long it takes a typical user to setup your product and get to proof of value but perhaps it can be done in 14 days, maybe even 7.

I personally believe that, in some cases and depending on the product, placing urgency on the user can yield positive outcomes. This also grants you additional freedoms to "nudge" the user, simple example, "setup an account and earn an extra week of trial".

Good work and good luck, keep experimenting and measuring.

I'd punt on offering free trials and instead offer a 30 or even 60 day money back guarantee. This is effectively the same thing but gets you access to money today as opposed to 30 days later. Learned from this terrific video from Jason Cohen: https://vimeo.com/74338272

Yep, seeing the effect of $.50 instead of $1 for trials when there are are lot more users & trials was really enlightening. I'll most likely do that. I could even not provide bank syncing for trials, or require a credit card. Will do something like that.

Definitely going annual as well - tempted to go only annual, but we'll see.

Annual is unappealing to most subscribers - but some percentage will go for it if it's available. Like 10%ish (may vary but definitely a minority). This minority will smooth your cash flow quite a bit.

What evidence/reason do you have for the claim that annual is unappealing to most? (I'm not advocating anything here, I'm hoping to learn something.) I would assume the same thing if there is no discount (no compensation for extra risk and time cost of money), but I can only guess at how people respond to annual options with various discounts. Do you have anything more (including your own guesses)?

One data point, for TeamPassword[1] we offer monthly and annual as described and about 1/3rd choose the annual plan. Would love to learn more data points to see how our mix is versus others.

[1] https://www.teampassword.com

Putting a high sticker price in front of someone is another one of those speed bumps to conversion. As the proprietor you'd probably prefer them to go for $70/yr, but $70 is a big difference as an impulse or low-deliberation purchase from $7. I'm the annual-plan-buying type, but I'm atypical in a bunch of ways - e.g., I live in a high cost-of-living area so while I'm not actually a money-disregarding rich person, my impulse-buy range goes a lot higher. (annual = 10x monthly is very commonplace)

Has anyone ever tried:

Annual plan only available after 6 months of membership?

While I understand that sounds counterintuitive, if not odd, it:

a) offers an incentive to stay.

b) with that goal line in mind, the offer of annual when getting there has special meaning. As opposed to a random "you've stayed. You must be happy. Wanna go annual?"

Again. I hear what you're saying, but in a nudge-y / Hooked sorta way it might work.

Bloomberg (magazine not terminal) did that for me, annual plan wasn't available during sign up (though there was a cheap 3 month trial); after a few months on the full monthly plan, offered me annual plan at a discount.

I took it.

Reply on news.ycombinator.com