Our pricing model was broken. Here's how we fixed it

This article, like most about pricing, misses the cardinal rule: price what the market will bear.<p>But they actually ended up in roughly the right place: free for unfunded services, a healthy monthly for lightly funded and a large monthly for well-funded.<p>They also get right that all substantive features should be offered in the free&#x2F;lower-priced tiers. The trick is to put &quot;enterprise-y&quot; features in the more expensive tiers. Best example is multi-user logins. This is not a core feature of the product but is a good proxy for a customer who has ability and desire to pay more.

Man, I hope they only included &quot;pricing based on cost&quot; for completeness. <i>Never</i> price based on cost! I mean, know what your costs are for each level of service, but if you are charging based solely on your cost, you&#x27;re not really doing &quot;pricing.&quot;<p>Setting your prices based on cost is like designing a product with features instead of benefits. Does HP design printers that can fold paper airplanes? No, they design printers that print business documents faster, crisper, and more reliably, because that is of value to businesses with money. In the same way, do they price ink cartridges based on the cost of the pigments? No, they charge by value, and it is worth a lot of value to the average business to get their documents and spreadsheets printed!<p>And don&#x27;t try to justify cost-based pricing by saying customers can just implement and host their own solution using open source code. Sure... and they can make their own printer ink and load it into recycled cartridges... but effectively zero percent of people will ever do that. :)

Ha! Interestingly, I just sent you guys a long mail about your pricing.<p>I think per unique visitor is relatively easier than per visit because an additional visit does not even tie to additional value.<p>The fact that you capture a lot of data from so many people (which costs you) does not mean that it brings value to me.<p>It is also important to remember that you are not prison in isolation but in comparison to your competitors. In your case, it is KISSmetrics and Mixpanel.<p>It is not something you can get perfect at one go. It has to keep evolving. You have a great product.

Would be interesting to see the effect of free trial length on their conversions (specifically shortening to 30 or 14 days and measuring the effects). At a company I work for, we found that shorter free trials actually led to more conversions.

I&#x27;ve been working on an open source alternative to proprietary analytics services, for those that want a middle territory between being a Mixpanel or Heap customer and having to roll their own.<p>Keep an eye out for Simonides if you&#x27;re interested.

Dan, you might want to do something about the logo. On the blog I saw it and read CHeap. not a radar + heap. or pie chart + heap. flipping the pie segment to another side would solve this.

I wish there was more information in the article about how it effected conversion rates.<p>They&#x27;ve created an interesting model here though.. Its basically free or one price, unless you&#x27;re really big. If you&#x27;re really big, they ask you to talk with a sales person, but clearly state a starting price (which I&#x27;ve seen few sites with a &quot;contact us&quot; do) to get the ball rolling.

This change makes your pricing a lot more feasible for the average web app, but I think it is still not perfect for many content sites whose traffic consists mostly of logged out short-term visitors.<p>Do you suggest sampling for those kind of sites?

Sad. I am the CEO of a social game with pretty big traction but with a pretty low user lifetime value (lower than in your example). With you I thought I could have a good analytics service at last. Not anymore.

Good stuff, I discounted Heap before because of the lack of a free tier and the lack of Android support, but it might be time to take another look.

Are new users (even those with less than 25k monthly) being converted to the new free tier?

So if you say you fixed it, do you have any before and after results?

A little off topic, but will Heap release an export API?