Pricing is hard. We need your help.

http://blog.keen.io/post/32269955563/pricing-is-hard-we-need-your-help

by dorkitude • 14 years ago

54 27 14 years ago

blog.keen.io

le_isms• 14 years ago

It would be nice to have an explanation of what Keen.io does in the article. Actually, even after browsing your site I still don't have any idea what you guys do, besides collect data and visualize it. There also doesn't seem to be any indication of -how- you guys do visualization. It's hard for me to think about pricing for your service when I have no idea what you offer.

dorkitude2• 14 years ago

Good point. We're not released yet, and there are probably < 100 people who have a good enough idea of what we're doing to really dig in and give us feedback

We'll edit the blog post to add a bit more background

update:

here's the working copy:

We make three kinds of APIs:

  -data collection APIs
  -data analysis APIs
  -data visualization APIs

For instance, if you had a social-local-mobile shopping app for the iPhone, you'd probably want to insert a rich event into Keen every time a user does one of the following actions: opens the app, does Facebook connect, likes/comments/shares an item, adds an item to their shopping cart, and completes a checkout. You send us this data using the collection APIs (https://keen.io/static/docs/data_collection/data_collection....)

Once your app is sending us stuff, your product manager may ask you to make her a little analytics dashboard, so she can agonize over it every morning. For instance, this dashboard could have answers to questions like "How many people opened the app each day over the course of the last week?" That question (and many way more advanced ones) can be answered using one of our analysis APIs (in this case, the Series API https://keen.io/static/docs/data_analysis/series.html)

Finally, suppose a few weeks later she's tired of staring at numbers and wants to see this information graphed visually in a line graph. That can be done using our visualization APIs (not yet released).

rgraham• 14 years ago

I think the what here might be less important than why for pricing. It's easy to get a sense of your platform from your website, but who is your target customer and why are they paying you? Does this save them time? Money? Make them money? Got any scenarios or customer sketches?

ljd2• 14 years ago

Because your customers are going to be more technical I would do the utility pricing. Pay per unit of usage.

If you need help with pricing, email me. Our company does algorithmic pricing via a REST API, I'll give Keen access for free.

To address your listed cons to utility style pricing:

1) Products aren't turned into a commodity because of how they are priced. They are commoditized when many perfect substitutions exist in the market.

2) I'm not sure if "not predictable" will go over well with your audience. Maybe you could use your own product to help people predict their bill with you. It would be a neat application of your software.

akavi• 14 years ago

You vastly underestimate the value of predictability in billing to a business's "Chief Fretting Officer".

If you asked companies if they'd rather pay a random amount between 100$ and 500$, or always pay 500$, 9 out of 10 would opt for the latter. And if you don't understand why, then you've never spent any amount of time pondering a business's cash flow.

JoshTriplett2• 14 years ago

If you asked them that way they'd opt for the former. However, most pricing structures don't provide such a clear-cut range; the edges feel quite fuzzy, and that uncertainty does indeed drive customers towards predictable-but-higher pricing.

ljd2• 14 years ago

My above statement is actually aligned with your sentiment. I think people use data analysis to help predict future events like how many customers you'll have or how much an expense will be.

I figured, since Keen is in the business of helping people with those predictions it would be a really awesome application of their own API. Customers could see their usage as part of the visualizing product and it would help them predict their bill.

codegeek2• 14 years ago

"pay a random amount between 100$ and 500$, or always pay 500$, 9 out of 10 would opt for the latter"

Why so ? You are perhaps referring to a business's ability to project/forecast cashflows which is critical but even then, if the number is always going to be less than or equal to 500, I would rather choose the first option.

bduerst• 14 years ago

Nope. You can use accounting to absorb the volatility.

runako• 14 years ago

I agree, but only because in practice the choice is usually a little different:

- always pay $500 vs. pay a random amount between $100 and $500, except when you have a really heavy month and you pay $3,000.

Utility pricing without the ability to set caps is problematic.

kirkers2• 14 years ago

The discussion here is exactly why we started the conversation.

Certainly, if we go this route, we'll make sure our users can readily monitor their usage so as to minimize surprises. Even still, we don't want our customers to have to kill the service in the middle of a billing cycle.

bromley• 14 years ago

And also the fear of an app (and consequently the API) getting hammered by the mobile equivalent of a clickbot (should such a thing exist), or some developer on the product accidentally programming in an infinite loop.

Oops, there goes thousands of dollars.

FWIW when I was recently involved in pricing an API, after much thought I decided the best option was tiered pricing with different rate limits (requests per hour) at each tier. At least with that structure you don't have to worry about angry customers with huge bills or customers bankrupting you with usage gone wild that they'll never be able to pay for.

The other risk of pure usage based pricing is that it attracts a cost-minimization approach to usage. That was my original plan but speaking to prospects put me off when I realized that many were willing to forgo all business sense and invest crazy amounts of development time to minimize API costs. And you still have to support them. (NB Yes you can have a base charge but many seem to think that inherently unfair when they have to pay for usage on top.)

lunaru• 14 years ago

I've always found Mailgun's pricing intuitive yet technical friendly. There's some tiering involved, but it's just a nice wrapper around utility pricing that doesn't feel like you're negotiating in cents.

jlarocco• 14 years ago

I think the description of what you do is just too vague to give any feedback on how to price it. "collect, analyze, and visualize" data could mean almost anything. I even browsed through the docs and couldn't figure it out.

Do you have a full blown example use case of how a customer would use keen.io?

The getting started page shows an example of inserting an "event" and counting it, but I'm not sure what that means, or what the end goal is, exactly.

dorkitude2• 14 years ago

We updated our description to make more sense and include an example

What do you think of it?

JoshTriplett2• 14 years ago

Whether you price by usage or have tiers, always do value pricing, not cost-plus pricing.

eldavido• 14 years ago

Agree; it's most in line with what the customer wants, facilitates segmentation in a way that's fair to both parties involved, and allows upselling when material changes in product or service take place.

Even if you use utility pricing, it's still best to frame it in terms of something meaningful to your customers.

jcampbell1• 14 years ago

Keen looks to be another Mixpanel. I looked at mixpanel's pricing, and it sis nicely structured into tiers so one can estimate quickly how much it will cost based on the size of the business.

Mixpanel's pricing is utility based under the hood (they automatically pro-rate or upgrade the plan based on what is cheaper), but the marketing copy is tier based. That looks to be a good approach.

waterside81• 14 years ago

Our pricing at Repustate.com is fixed fee for a fixed quota of API calls. Usage-based billing sucks because of the unpredictability on both sides, plus it makes invoices much easier for us. Like others have mentioned here, predictable costs are good.

reubenswartz• 14 years ago

I've helped a number of companies put together pricing plans, and the first thing I usually say is "talk to your customers", so it's great that you are already doing that. However, the real question is not "how should I price this?" but "why is this awesome to you? (what is the value?)" You should absolutely price based on value-- if your value is greater than your cost, you can win. If not, you need to fix something. If you don't know what that value is, spend your time there, and the pricing will flow from that. As mentioned by others, you can always grandfather existing customers if you decide to change.

(That said, I'm guessing that simple tiers + custom enterprise plans if needed will be the way to go. The trick is figuring out how to set them up.)

Pricing is hard. We need your help.

Related Stories