This is a very informal, off-the-cuff survey of some web applications in the Rails ecosystem that have varying payment plans, how they present them, and some of the differences among them.  The idea is not to compare the services, but to start to get a handle on how to price services on the web.

Background

I’m working on launching a small, very niche Rails-based web application in the near future, and am looking to charge a small amount per-month for it.  Mostly as a way to make ends meet while I’m unemployed, but also because I’d like to provide a few extra features to those users who already donate to the free JBidwatcher application, and because the web application uses S3 and I have to defray those costs.  While I’m very grateful my users are contributing because JBidwatcher itself is useful to them, I’d like to offer a ‘premium’ value both for folks who have already donated, and for people who don’t feel comfortable just sending money, and want to see something extra for it.

This led me to be curious about pricing.  I pay for several web applications already, and expect that eventually I’ll have to pay for more if Google starts actually asking for money.   I decided to dig into the applications that are out there, and see what kind of pricing models I should be looking at.

From my informal survey, a lot of the subscription plan models used by Rails-oriented companies naturally grew out of 37signals and their fanatic devotion to the concept that web applications should be good enough to charge for, and should be charged for.

The products I’ve looked at for this are Lighthouse, Tender, GitHub, Highrise, Basecamp, Campfire, NewRelic RPM, CrazyEgg, Hoptoad, Thunder Thimble, and Pingdom.  I’m sure there are many others out there, but those are the ones I deal with on a semi-regular basis.  I have paid accounts with GitHub and Pingdom, and have used (at other companies) paid accounts on Hoptoad and NewRelic RPM.

One specific attribute of all of these applications is that they have varying levels of paid accounts, not just a free/paid[/lifetime] like many other services (LiveJournal, Flickr, LibraryThing, to pick a few that I personally use).  I like the spectrum approach as it gives users a choice as to how much services they’ll need, and what the amount they’re willing to pay for it is.

Raw Pricing Plans

First I want to present the ‘raw data’, a link to the plans and a brief table with the plan names and the monthly cost.  Then I’ll point out a few things I found interesting about the data.

So what’s interesting here?

Lighthouse, Tender, GitHub, Highrise, Basecamp and Campfire all differentiate on disk space used.

Lighthouse, Tender, GitHub, Highrise, Campfire, Hoptoad and Thunder Thimble all have some variation on the concept of ‘user’ that you can pay more for more of.

In everything except NewRelic and to a lesser extent Hoptoad, all the capabilities of each application are available to all user levels, just at varying quantities.  Feature distinction only exists in those two applications.  In Hoptoads case, the distinction is between free/non-free; if you’re paying, you get all the services.  This leaves NewRelic as the only one that deeply distinguishes between features available to different paid levels.

GitHub, Highrise, Campfire, and Hoptoad all have SSL support as an ‘add-on’ feature; it’s not part of the free accounts, and in some cases it’s not part of the basic level paid accounts either.

Discounts

When I signed up for Pingdom, they sent me a ‘70% off the first year’ invitation, which reduced the price to roughly $3/mo. for all the basic plan amenities; presumably under the theory that they will be able to re-subscribe me in a year at full price.

NewRelic is running a 20% off discount currently, which lasts as long as you have a paid account.

Pricing Plan Layout

Historically I seem to recall most of these sites had the table style of comparison chart (still used by NewRelic, Hoptoad, Thunder Thimble, and CrazyEgg), but most have converted to the ‘box’ style of comparisons.  GitHub still has the table when you view the plan chart from your account settings, but other than that Lighthouse, Tender, GitHub, all the 37signals products, and Pingdom are using separate boxes for each plan level.

The 37signals products and Pingdom use an outsized box to emphasize one of the plans, presumably to drive signups to that plan.  CrazyEgg does the same thing within their table style, also.

I kept the order the various products displayed their prices, and it’s noteworthy that all the 37signals products, CrazyEgg and Lighthouse start with the larger plans on the left, and go down from there.

It’s also interesting that all the ones marked as ‘and a free plan’ de-emphasize the free plan by putting it in small text under the large table of paid options.  Two of them, Tender and Thunder Thimble offer a 30 day free trial, but no ongoing free plan.

GitHub and NewRelic are the only ones whose plan details go below the fold.  GitHub’s plan upgrade doesn’t, but their new sign-up plan list does.

Disk Usage

So for the applications which differentiate on disk usage, how much does $50/mo. get you?

This pattern is roughly the same for the $24 price point for all of them. 500MB for Lighthouse and Tender, 2.4GB for GitHub, and 3GB for the 37signals products. This suggests that limiting on disk usage ranges from 20MB-200MB per dollar per month, a pretty wide range. Estimating based on S3 costs suggests a per dollar per month storage amount of around 2.3GB, but that relies on one upload, one download, and ongoing storage. If your usage is asymmetric, or storage is temporary, the S3 based cost can vary a lot.

Users

For the applications which vary pricing based on users, how much does around $24/mo. get you (I chose this number since Hoptoad doesn’t have a $50 price point)?

I believe this is a wide range mainly because per-user data is usually relatively light, and so it has more to do with how complex the application’s interaction between users is, rather than a real per-user cost. Still, some numbers do come out of this. The cost/user ranges from $0.78 to $4. At the $50 price point, the cost/user for qualifying apps is $0.82 to $3.27.

Summary

I don’t pretend to know pricing, or sales, but I like to believe that in aggregate the people putting these sites together do.  I see a pretty good argument here for feature parity among price points, but finding quantities that can vary between prices.  There is a clear value to users and disk space used, so those are early things to look at when pricing an application.  SSL support is a common feature of paid plans, and not of free plans.

There’s a definite movement towards boxed plan details, over tabular feature comparisons.  Ongoing free plans still exist in the majority of applications, but are de-emphasized in most, guiding users towards the paid plans.  Overall, the plans are simple, only falling below the fold in two cases, and relatively easily consumable in all.

The lowest payment point plans are $5-40, with a bare majority falling in the $5-12 range, and all the rest but NewRelic falling in the $19-$25 range.

Closing

I hope this has been an interesting and potentially useful survey of a few pricing plans for applications generally in the Rails ecosystem.  Any mistakes are mine, and I’d very much like to hear about them so I can fix them.  Other data points are welcome, and points I might have missed that would be valuable to folks thinking about pricing are welcome, and even encouraged!

I did this for my own edification, but I’d also very much like to know if others find it interesting!

Best of luck, and may figuring out pricing not be as much of a pain for you as it is for me!

–  Morgan Schweers, CyberFOX!

A concerned user recently asked me how I was doing in the aftermath of the issue with eBay sales of JBidwatcher, specifically:

You seemed pretty depressed about it in your post to the website.

I was.

There was a really bad week there, while I was dealing with all of it, back and forth, and just feeling like crap. I got a lot of user feedback, from a LOT of people, that reminded me, as the concerned user put it, not to let the few jerks make me give up.

I took some time to work on other projects, and I’ve been fiddling with the next major rev of JBidwatcher, mostly cleaning up the code, improving the source layout, fixing small things that nobody else will likely ever see, and writing silly features just for the fun of it. (Like making the internal webserver take ‘events’ to be posted to the various subsystems, so you can add an item, do a bid, or even tell it to fire off a sound effect through a REST-ish interface. You could theoretically ’script’ JBidwatcher through that.)

I’ve also mostly moved the code base of JBidwatcher to Java 1.5 (mmmm, tasty generics!), since 1.6 is now out. I’ve also been experimenting with including ‘Derby’, an embeddable (in the ’ship with program’ sense) tiny SQL-based database, so that JBidwatcher’s memory usage doesn’t grow at the same rate as the number of auctions. Also so that it can offload completed auctions, so they’re not kept in memory anymore at all. Yet another thing I’ve been playing with is including a scripting language (something simple) which would get run on certain events, which would allow for making some of the complex rules people have requested as features. I’ve also written up an FAQ I need to publish on the site. (The first question addresses my inability to answer emails consistently, in fact!)

Anyway, all told I (and work!) have been keeping myself busy, albeit quiet. It all helps me get past the issue with the people selling JBidwatcher. Future versions will probably not be open source, however. I may expose the source, or open certain sections, but almost all open source licenses explicitly allow what those folks were doing, and I’ve determined that it’s beyond what I’m comfortable with. One of the things people repeatedly said in private emails was that the open source nature of JBidwatcher was not critical to their appreciation of JBidwatcher. This means I’ll need to extricate myself from Sourceforge in various ways, and cover my own purchase of IntelliJ IDEA, but I think donations will have covered that.

I don’t want to charge for JBidwatcher; I prefer people using it and deciding for themselves what it’s worth to them. Plus, because it’s scraping eBay, I feel bad about asking for money for something that could break the next day. So I expect the program will continue to be no cost. I’m thrilled to get donations, of course, but I don’t build JBidwatcher to make money; I have a day job for that.

At the same time I’m working on building other projects so my morale won’t get torpedoed so badly when someone messes with the sole project I’ve been working on.

One of the other projects I’m working on are a health tracking tool (weight, blood pressure, hours slept, water drank, steps taken, foods eaten with nutrition information, and more stuff like that, with pretty graphs and sparklines (my weight trend: )). The other major one is a comprehensive multi-user outliner tool. Both are entirely web based applications, unlike JBidwatcher, and both are in Ruby on Rails.

I’m sorry that I haven’t been dedicating more time to JBidwatcher, but it’s been fundamentally working okay recently, and I needed to blow off steam by doing cool new stuff. So I’ve been letting it percolate, and rekindling my coding passion by working on other interesting problems.

That’s the status as of now; I hope that this sheds some light on my thought processes, and what I see in the future for JBidwatcher. One important thing to take away is that YES, there is a future for JBidwatcher.

Thank you, every one, who wrote me, donated, or just thought well of me during all this. I appreciate it a great deal more than I can express.

– Morgan Schweers, CyberFOX!

I’ll pass along a piece of advice for anyone looking to start a company, from someone (me) who’s been through several startups.  (McAfee Associates and PayPal being two very notable, very successful, and very different, ones, in different decades.)  The advice is useless ever since 1994 because the ‘free money’ vibe of the VCs has infused the business world and made it hard to follow, but it’s free advice anyway, and worth whatcha pay for it. 

If you’re going to start a company, be profitable FIRST, before you ever talk to VCs.  If you don’t NEED their money, but you’re on one of the beautiful adoption curves, they’ll fight each other to be the one who gets the right to give you money.  Sure, you don’t need it, but if spent right, it can (1) pay for the dog’n'pony ‘going public’ show, and (2) move you one or two rungs up the doubling curve.

The funny thing to me is that people take the money out of the get-go, to move them up 1-2 rungs when the power curve is low, boosting them from 1000 users to 4000 users.  What they should (in my oh-so-not-humble opinion) do is grow slower for a little while, then take the money when jumping the power curve would mean the difference between 250,000 users and 1 million users AND you’re not dependent on the VCs for your existence.

This means growing slower at the start, only growing as fast as you can afford to, keeping yourself as close to cash-neutral as you can, and always being one cost-cutting exercise from being profitable.  It’s not the preferred method for a class of folk who were raised and bred on tales of overnight millionaires, and ‘get big fast’ (which only works for a very SMALL subset of companies) but it’s the way to build a sustainable, exit-strategy-free business.  (Exit-strategy-free meaning you don’t NEED one, not that you don’t HAVE one.)

I saw it done at McAfee Associates, and we completely dominated the conversation with the VCs.  It wasn’t a request for money, it was a bidding war on the part of some top-notch firms, and that was in 1992, before the first bubble even started to be blown.  We were chugging in about $10Mil/year as I recall, and spending less than a quarter of that.  I don’t believe that company was EVER unprofitable, from the quarter it was founded.

McAfee Associates took the company from “the three Fs” (family, friends, and fools) to profitable without really going through the intervening step of angels or venture capitals.  (Now, for what it’s worth, some of those 3Fs were…understandably upset that they didn’t see any return when the company made it big, but that’s just part of the darker side of the history of McAfee Associates, and not directly relevant to the point I’m making.)

Now McAfee Associate’s distribution model was nearly free (distributed by BBS), advertising was ‘make a good product and get people to talk about it and show its use off’, the product was…well, you could joke that it was viral.   People used it, found it was useful and necessary, and handed it to other people who probably needed it.  The product was free and fully functional, but every run it put up a message, basically that ‘if you’re using this in a company, educational institution, or government organization, you must purchase a license’.  End users paid the registration fee sometimes, which was a nice base source of income, but companies leapt to license.  Mostly because their employees were using the software, needed the software, and were handing it around inside the company.  Someone would point out the license issue, and the company would call us to pony up.  We never had to call anyone to sell the software, it sold itself, and our users sold it for us.  Most people wouldn’t imagine it, but companies really are decently minded when it comes to dealing with other companies.  Or at least so afraid of lawsuits that they’re willing to be decent citizens.  Sometimes a big company would use their size to pressure us to give them a discount, which we were happy to do, because it was nearly free money anyway.  Our overhead was fixed, we didn’t advertise, we told them to download themselves a copy each quarter, or if they paid enough we’d send them a floppy once a quarter.
What’s changed in the world since then?  Well, a lot, but I’ll put forward that mainly the scale has changed.  Free distribution through BBSes reached a large percentage of the computer using population back then.  Distribution through the Internet reaches a HUGE percentage of the computer-using population now.  If you make a good product, people can talk much more widely about it with blogs.  Companies are still looking to be good citizens, primarily, and there are a LOT more of them out there.  Oh, and you probably don’t have to mail them a floppy each quarter, no matter how much they pay.
So the one of the biggest things I’d suggest to ’start profitable’ is to target your product at corporations, but make it attractive to end-users as well, so they’ll bring it into the company.  As Willie Sutton didn’t say but is said to have said, when asked about why he robbed banks, “That’s where the money is.”

–  Morgan Schweers, CyberFOX!

So I’m very, very much looking forward to Gnomedex, coming up at the end of this month. I went last year, and it was really cool. I went to the Seattle MindCamp, which was intense, a lot of fun, and tiring. I met good people at both, although the MindCamp folks were more…relaxed, I think. The knowledge sharing was much more peer-to-peer at MindCamp, and I really reveled in it.

Gnomedex is looking to be even more interesting than the last one, with the addition of Senator John Edwards as keynote speaker. I don’t expect much from him in particular, but the added attention it’ll give the conference should be interesting.

More than anything else, I’m looking for the little moments, like last year when I watched the JotLive Wiki updates stuff happening as the guys debugged it while sitting in the back row of the conference. In retrospect, it presaged the Google Spreadsheets interactive/shared updating. The big presentations on stage generally came to nothing. The little moments, and the side discussions are where the action was, and I expect that to be true this time as well.

Plus I get to show off my new house to some friends who are coming into town for the conference!

I need to get my act together, get some software updates out there, and psych myself up for the conference. I’m not a big name guy, so my contributions tend towards having useful pieces of technology to show off and play with.

On a more depressing note, my mother asked for money again, which coincides with the start of my paying mortgage payments, NOT a good time for this. However, I think I may re-release JBidwatcher with the affiliate program enabled again, and funnel the money from it to help her out. I could simply sell the program, but without being able to use the eBay API I can’t promise users it’ll be consistently functional, and that seems like a sucky thing to do to paying customers. At the same time, the program drives revenue to eBay, so it should be of value to them.
This is yet another place that I can get some help at Gnomedex, not so much directly for my mother, but to get an idea of ways that people have monetized their passions outside of work, in order to get an extra income when necessary.

My goal at this years Gnomedex is to keep my ears open; to hear the rumblings of what will make next years tech news, when writ larger. I need to keep in mind the small conversations, and imagine them happening in the halls at Google, or Microsoft, and imagine what they’d do with them…

So many people probably think that Gnomedex will be all about Podcasting, and VLogging. I think that’s just the cover story. The medium(s) are not the message. I think it’s about new ideas, new directions, smart people, and listening carefully to what people are showing off.

Yes, a dozen video hosting sites are making history, and providing all sorts of interesting value, as well as amusing video. Yes, podcasts are so mainstream they’re available on iTunes Music Store, CBS has turned over its incredibly valuable properties into $1.99 iPod Videos, and Suicide Girls post iPod Video-compatible introductions. Perhaps the abundance of new video sources will push out the television, and eventually replace it. This actually interests me less, because it’s a Big Media world.
I’m more interested in what people are saying, rather than how they are saying it. I don’t so much care about the media they use, I care about the ideas they’re getting across.

I probably won’t vlog, I’m not photogenic enough. I don’t really podcast, for the same reason, nobody wants to hear my ideas interspersed with ‘umm’, and such. If anyone is interested at all, they want the thoughts filtered, cleaned up, parsed, and posted, so they can be searched, summarized, stored, and (most often) skipped.

If there’s any serious value out there, in interesting but HARD technology, it’d be in automatically extracting the text of a voice discussion, describing the scene in a video, etc., so it’s searchable using text search engines. This is a Hard Problem, but it would suddenly make the world of VLogs, Podcasts, and even TV and movies, all searchable, indexable, and accessible to everyone.

There’s a world-changing technology, if ever there was one.

See? I’m already getting my mind into the mode for the Gnomedex…

– Morgan Schweers, CyberFOX!