3 ways to make money with your software - Signal vs. Noise (by 37signals)

3 ways to make money with your software Jason 14 Aug 2006

43 comments Latest by Martin

There are three primary ways to generate revenue from web-based software. Let’s take a look…

Advertising-supported
Give your product away for free and sell advertising around it. You can do this with desktop software (as IM clients usually do) or web-based software.

It’s definitely the easy option for web-based software: Just paste in the Google Adwords include and bam, you’ve got ads. But pasting in ads doesn’t mean you’re pasting in revenue. Unless you have massive traffic, revenue can be very hard to come by using this strategy.

This option also doesn’t work for password-protected, log-in-required products since Google can’t get behind your authentication. If you want ads on a password-protected app, you’ll need to sell them yourself — and that’s a full-time job.

Grade: B-. Personally I don’t like this model because it throws your priorities off — you’re beholden to your advertisers and their screen real-estate demands instead of your customers. It’s tough to serve these two groups at once since their priorities are likely to be different; The advertisers want to win your customers’ attention but customers want to focus on the task at hand. A true win-win is tough in this scenario. Possible, but tough.

Subscription-supported (or single price)
This option asks the customer to pay for the service. This model is really saying, “We think our product is worth paying for and if you agree we’d love to have you as our customer.” It can be a monthly/yearly subscription or a single one-time price like traditional software. From a revenue standpoint, the subscription model really fits. If you look at in reverse, your company is subscribing to each customer’s bank account (not in an evil way, of course).

The downside is that it’s tough to make a product that people will pay for. Some customers are willing to pay but are there enough of them to keep you afloat? That’s part of the territory when you sell anything though. However, if enough of the right customers find you, healthy revenues await. Plus the people who pay you and the people who use the site are the same. That helps bring priorities in line. When there’s a clear connection between customers and revenue, it’s a lot easier to focus.

Grade: A. I recommend this model. If you go into your project thinking you have to build something worth paying for, then you’ll likely build something worth paying for. You’ll try harder because you know it has to be really worth something to people. Of course, even good ideas sometimes fail. But building-to-charge puts you in the right mindset to succeed.

One more thing I’ll say about this: When people pay for your product they have a vested interest in seeing it succeed. They want their investment to be a good one. They want to believe in it and care for it. And they want you to stick around. All that helps ignite passion and interest around what you’re doing and that’s a very good thing.

Support-supported
This is the strategy of the open-source software movement. SugarCRM is a good example of this. Their software is free — but if you want vendor-backed support help you’ll have to pay for it.

Grade: C+. On the surface it seems like a good deal, but for who? If the revenue model depends on helping people, then there’s an implicit motivation to make things difficult instead of simplifying and make things easier. This model punishes tools for being intuitive and easy to use. In essence, you wind up selling complexity and a secret language that only insiders can speak. That’s good for the company but bad for the customer.

A combo meal?
There are other ways to make money off software too. You could create a model that mashes up the ones mentioned above, Or you could license your software to other companies to sell or give away. Or something completely different.

The bottom line
In the end, we think building a real product that people are willing to pay for is the best way to go. It’s a dive with a high degree of difficulty, but if you can pull it off you’ll be swimming in profitable, sustainable waters.

Related: Russell makes sense this time

43 comments so far (Jump to latest)

Ohad Lutzky 14 Aug 06

If you intended to make money off Rails, keeping it GPL’d, which of these three methods would you use?

Darrel 14 Aug 06

“On the surface it seems like a good deal, but for who? If the revenue model depends on helping people, then there�s an implicit motivation to make things difficult instead”

While that seems plausible, I don’t think it’s really true in practice. No developer wants to purposely make something more difficult…just for their own sake. And the more difficult it is to use, the more time you’re suporting the people, which takes away time doing other work.

I think most people (ESPECIALLY IT departments) don’t really want to USE support, they just want the knowledge that it’s there if they ever do want it. And those companies that offer support only packages certainly don’t want the product to purposely be complex, as they just want people paying for support, not actually using it.

You also forgot the 4th model of making money with software: Make sure all of your releases are full of bugs and incomplete features, ensuring that you can always get people paying for the upgrades just to get the bug fixes. I’ve found this is the model that most CMS vendors prefer. ;o)

JF 14 Aug 06

If you intended to make money off Rails, keeping it GPL�d, which of these three methods would you use?

You could write books, conduct workshops, create a job board, etc.

Joshwa 14 Aug 06

On the other hand, subscription is a little trickier when your product relies on ‘network effect’. Then, you’ve got to make sure you give enough away for free that folks sign up and you build the network, but keep enough behind the pay-door to compel people to fork over the dough. It’s a fine line to walk…

Jim Greer 14 Aug 06

A good analysis - the combo method can work very well in some fields. For instance, I used to work for Pogo, the web games portal that’s part of Electronic Arts. Pogo started as an ad-driven service, and launched a subscription product in 2003 - it now has 1.4M subscribers at $35/year. That plus ad revenue and microtransactions makes it a star division at EA (web games are a lot cheaper to make than Madden ‘07 and Spore).

Yahoo Games did the same thing with Yahoo All Stars - they have very few subscribers. But their site is a lot worse, and depends on the firehose of traffic Yahoo provides.

My new web games company is going to combine ads with microtransactions for premium features. We’ll probably roll some of the microtx features into a subscription service once we figure out what people like best.

Berislav Lopac 14 Aug 06

As Jim mentioned, micropayments are a fourth way, but it’s not suited to every project.

Nathan Bowers 14 Aug 06

“Support-supported” software can go either way. For example, MySQL is an excellent and easy to use product with a training and consulting revenue model. MySQL makes their product easy to use because what they “lose” in consulting fees is more than offset by ubiquity.

On the other hand, IBM sells support for open source software, but they also sell lots of extras to their enterprise clients. These “extras” are usually a nightmare, and they require many IBM consultant hours to set up and troubleshoot. I’ve seen this first hand with IBM’s Websphere and the raft of junk they sell with it.

Yes, the revenue model can tell you some things about software quality, but you also have to consider who’s buying. If you’re selling software to Fortune 500 IT executives (who have every incentive to keep their head counts and budgets bloated) then it’s in everyone’s best interests to ship software that’s hard to setup, use, and maintain.

Chris Harrington 14 Aug 06

“�On the surface it seems like a good deal, but for who? If the revenue model depends on helping people, then there�s an implicit motivation to make things difficult instead�”

I disagree with this, especially with reagrds to SugarCRM. Vendors such as this aren’t motivated to make things more difficult, they are motivated to add in features and functionality to make the product worth paying for. There is more to that model than just charging for support of an open source project. Many GPL projects, especially the vendor backed ones like SugarCRM, Snort and Asterisk, have robust communities and mechanisms for free support. I would concede that there *may* be motiviations to not make the GPL version any easier. That is different than intentionally doing so.

—Chris

Daniel Haran 14 Aug 06

I had to choose advertising. I’m hoping ads will increase the use and usefulness of the site.

It can be difficult to decide what will be behind the pay-wall, and achieving a network effect without a blog like this one can be tough :)

Adam 14 Aug 06

You could write books, conduct workshops,…

Is this really that much different than selling support? I’ll agree its a better implementation than your example in the post, but it seems to fall under the same category, no?

Ted Shelton 14 Aug 06

Call me crazy, but isn’t the right strategy very dependent upon the kind of product and the demographic? For example, if you want to go after the high school MySpace demographic, I think the subscription supported would fail - they don’t even have credit cards, how are they going to pay? Yet there are advertisers that want to reach these kids. First, determine your demographic, take into account the type of product and how it will be used, THEN determine your business model…

Deryck Hodge 14 Aug 06

The idea that open source companies purposely make products complicated to advance support subscriptions is ludicrous. You’re advancing a theory based solely off your own conclusions, rather than any particular evidence. Your conclusion might seem likely, but I don’t know of any company that consciously or subconciously does this.

Now, some open source products may have usability issues, I’ll give you that. But’s it’s largely a result of geeks designing for geeks, or tech companies not understanding end users, rather than any commercial motivations to keep software complex and unintutive. It’s your blog, your right to advance whatever opinion you have, but please don’t accuse an entire industry (and set of developers) just because you have an affinity for one business model.

Deryck - He didn’t say that “open source companies purposely make products complicated”. He said that selling support puts the incentive in the wrong place. It doesn’t mean that every company succumbs to that incenctive.

Aaron Kassover 14 Aug 06

Don’t forget the fourth option: Give it away for free and hope to get snatched up for millions by Google or Yahoo!.

Jim, if you’re going to picky about word choice, then quote me accurately, too. :-)

I said “the *idea* that open source companies purposely make products complicated […]” I’m criticising the conclusion more generally not Jason’s exact words, I believe.

ML 14 Aug 06

The idea that open source companies purposely make products complicated to advance support subscriptions is ludicrous.

Deryck, we’re not arguing that any specific open source companies purposely make their products difficult to use.

We’re talking about the support model as a concept. There’s less motivation for a company to make a product easy to use since it gets paid to provide support for that product. We prefer a model that unites revenue with ease of use instead of placing those two things at odds with each other.

ML, I don’t disagree with you that a subscription-based model is “a model that unites revenue with ease of use.” I just disagree that an effect of a support-based model is to lessen the motivation for a simple design.

If I sell a Linux server that has to play nice in a mixed network, that server will have to be setup to work with Active Directory, OpenLDAP, Kerberos, etc. It may have a web server, a file and print server, etc. Some things really are complicated and selling support adds value. There is still the desire to make the design as simple as possible because you’re already working in a complicated space.

I’m citing a specific example where the model holds true, IMO, and the scenario most closely tied to “the strategy of the open-source software movement.” Now if you were thinking about some hypotetical open source web application business model, then that’s a different beast. But that’s hard from me to know from this post. Specifics are always better than broad generalizations.

Some things really are complicated and selling support adds value.

The post was about about “web-based software” not complex infrastructure or server implementations. We’re talking about end-user software, not low-level infrastructure which is inherently complex. I’m sorry if we didn’t make that clear.

Probably the wrong place to post but how come you guys are not speaking here

We just didn’t feel like it was the right event for us to be part of.

commentKing 14 Aug 06

Jason, from what I read, these are all B2C models, what happened to B2B. A very profitable and common method of generating is selling non-exclusive licenses (or exclusive if the price is right) to other business. There are quite a few companies doing this right now, including one that got acquired by AOL this morning. :D

Anonymous Coward 14 Aug 06

Or you could license your software to other companies to sell or give away.

Said.

Jason, thanks for the reply. I understood you were speaking about web software. It’s just when you used the term “open source software” to describe the support model, the post read, to me at least, like you meant the model as more widely used was somehow a bad one.

I guess, too, I just don’t associate that model with web apps much, since so few really use it. I would agree that it’s not a good model for a web app, though truthfully, not as much because the model provides no incentive for developing ease of use, but because it just doesn’t make sense for a web app. What is there to support? :-)

Anyway, thanks for clarifying.

Mike 15 Aug 06

“It�s definitely the easy option for web-based software: Just paste in the Google Adwords include and bam, you�ve got ads.”

I think you meant Adsense not Adwords.

Good read, but thought you might want to correct that.

Ed Byrne 15 Aug 06

SugarCRM is not ‘support supported’ - it’s over 1,000 dollars for a five-user license of the professional version, which is the feature-rich one any business serious about CRM has to go for (the others don’t have pipeline tracking).

Not sure how that makes it open source - sure I get the source code and can edit it, but I’ve still paid for it.

Ismo Ruotsalainen 15 Aug 06

We do custom webapps for our customers. Our pricing is simple. We charge only for project (like intranet, website or webshop) and there is no hosting- or using fees.

If our customer wants, we love’d to make a “development contract” (i dont know is term right in english). Anyway the “development contract” means that we use some time every month to updating, marketing and developing clients webshop. Client can deside how much time we use, for example 2 hour or 15 hours / month.

We think there is one thing that sucks on hosting-fees: webshop doesn’t get better. Thats why we use “development contract”: Webshop is getting better every month.

Our development projects based on customers and visitors needs, site statistics and conversions. We design and code what site/shop really needs.

Bottom line:
We dont have to sell more and more every month. We can sell “option removing” if site/shop need that.
We are happy with this pricing model and our customers love’d it too.

Sebhelyesfarku 15 Aug 06

I have an ad filter usercontent.css in Firefox, nobody will make money on me from advertising!

Dan Boland 15 Aug 06

Good post, the best from you guys in recent weeks.

DAR 15 Aug 06

I think it’s been pretty clear from the very beginning of the Net “revolution” that users *don’t want* to pay for things, and prefer advertising supported. Whether you like it or not subscription-supported (or single price) gets a grade D, while ad supported gets an A.

I realize that you don’t *want* that to be the case, but nevertheless it is. Pretending that it’s otherwise doesn’t make it so. The market has spoken and people want things on the Net for free. So much so that companies and VC’s these days look at charging money primarily as an *impediment* to gaining users.

Hey - it’s up to you, but IMO you ignore this reality at your own peril.

Ilija Studen 15 Aug 06

Dar, one problem is that they are not pretending. 37S have a good, sustainable business model - something that most of the “Web 2.0” businesses don’t have.

Its not a sin to ask money on the web when you think that you provided a good product (or service). There are situations when this is not the best way to go, but that does not mean that everything on the web need to be free or ad supported.

Great post btw.

Hui Zhou 15 Aug 06

“On the surface it seems like a good deal, but for who? If the revenue model depends on helping people, then there’s an implicit motivation to make things difficult instead”

The current software seems exclusively focused on serving the general public and insist on usability for general public. But people diversifies from each other. The very reason we cursing some of the UI design is exactly because the designer tried too hard on guessing people’s intuition. Use editor for example, vim and latex is very intuitive to me, and word just feel awkard. And use media center as an example, I never watch live TV and use epg, how would the desinger guess my intuition?

I would like to see a future trend that software are geared toward cusomization. Of course there will always be basic versions for general public dummies, but there will be great need from geeks, professionals, hobbiests that can’t live without customization. By then, I would hope to see a new lineup of software developers that can do those on demand customization for a living.

The current software model is not fit for such profession. Customization is way too hard and expensive. To customise a browser, one need figure out every bit code. But I hope in the future, there will be a mjor software industry providing standard basic software parts — just like nut and bolts, simple, basic, useful in many place, cross platforms. And I expect the software support profession would be as prosperous as the autoshop these days.

Morten 15 Aug 06

I’m all for the subscription supported with recurring billing. Can anyone recommend a good partner for handling the transactions, credit card validations and so forth?

Julio Nobrega 15 Aug 06

You can show AdSense on pages that require a login, just give access to Google’s MediaBot (by agent name and/or ip).

There’s also affiliate links. A project management app can sell project management books and courses.

Mitch & Murray (from Downtown) 16 Aug 06

I second Morten’s question - who is a good partner to work with for all the back office credit card transaction that a recurring billing business model requires?

Sunil Tanna 16 Aug 06

I have to wonder about your grades for Advertising supported vs Subscription supported.

To be fair, it does depend a lot on the app.

*But*, people (i’m talking about the great mass rather than a few elite rich folks) will put up with an awful lot of intrusive advertising, in order not to pay.

For example - how many people would pay for web search?

Prior to the advent of google, the search engines were going to the more and more intrusive advertising model. It wasn’t uncommon to see a major search engine with sponsored links, banners, and pop-ups all on one page.

Post-google, the other search engines have toned down the intrusiveness of ads as part of their competition to gain users, but many ads are still there, and google’s have infact gradually grown more intrusive.

And I can’t imagine anybody (well virtually nobody) paying to be rid of the ads.

Ian McAllister 17 Aug 06

I couldn’t agree more about your point that people will pay for software that provides value. I think you’ll create a better product by focusing your effort on 10,000 paying customers that really use your product then 1,000,000 where only a small fraction occasionally click on ads.

What about fees, either posting or transaction? Amazon, ebay, craigslist and others seem to make money with this strategy. Goals line up pretty well here too since your site needs to provide value for your customers before they’ll pay.

The fee model is also hard, but for the right reasons. I think the barrier to entry here is higher because marketplace-type sites need to get some kind of critical mass before substantial value for the customer is created.

Morgan Schweers 17 Aug 06

Greetings,
Interesting and timely, for me. It’s a great article, but I might be saying that because it meshes with my own mindset on it.

I ran the numbers for myself, trying to make exactly this decision for a new webapp I’m developing.

Baseline
On a site I run for a reasonably popular open source app I’ve developed, I make .32 cents (that’s just under one third of a cent, not 32 cents) per visitor via AdSense. It would take 1 Million visitors a month to make $3200/mo. This is a useful number for me, because it approximates my mortgage. :)

It’s REALLY hard to get to a million visitors. Right now, my AdSense-enabled site more or less pays for its own hosting, which is all I ask of it. I want more from my new website/app, though, and a million visitors is just a pipe-dream to me.

If I charged $10/*year*, I’d need to produce something that was of value to just shy of 4,000 users out of the hundreds of millions of users out there, in order to make the same amount on average.

Numbers
With roughly 700 million internet users, being of monetary value to 1 person per 175,000, nets 4000 users. (Ignoring issues of discovery for now.) In order to derive roughly the same money through ads, I need to be of constant (monthly!) low-level interest to 1 person per 700, a much harder bar to meet at 250x the number of users, imo.

Eggs-per-Basket
Also, if I have 4000 users, paying yearly, it would take all of them going away to remove my income (and you better believe I’d start making changes before they ALL went away!). If I’m ad-based, it takes a minor Google change to eliminate some, and maybe even all of my income.

Basic question
I guess another question for each person, in choosing what to build, is which is easier for you, build something good that is of value to a niche audience, or drive constant, high levels of traffic to your site?

As a software developer, I find the former is far more to my tastes.

— Morgan Schweers

Martin 26 Aug 06

Many good points here, but you should mention in this blogpost that you are partial in this matter.