The cost of free software


We've set a precedent as consumers by choosing "free" products over paid products. This fosters an environment that caters to startups seeking acquisitions and data collection, rather than dollars directly from their users. It makes for bad products and short-term thinking.

Nothing is free

When the average pay of a San Francisco engineer is $103k, you quickly realize how much it costs to get a simple idea from design to production. Tech startups do have the benefit of usually requiring less physical resources, but engineers are costly and in short supply. Some startups also require more than just engineers to get the product rolling. Depending on what you are building, it may take 10 or 20 sales people to build strategic relationships.

So how do startups afford these employees if customers are not directly paying? How could so many startups offer something for free when they are burning through so much cash?

It's simple. They're spending money, just not the money you think.

The four primary business models

These are four sections representing the different ways startups make money. They're color coded from green (good for users) to red (bad for users), with some products acting as hybrids between two areas.

Posthaven (the platform that is used for this blog) is a pay service, meaning they make all of their money from their users. The more users, the more money they make. If they operate their business optimally, they'll theoretically keep making more money as they add more users, hiring necessary employees as the product scales. This is the pinnacle of "user focused" startup. They want you to pay, so they'll cater the business around your needs. Fastmail, a service I use for email, also is a direct pay product.

Most blogs operate within the "Ads/Sponsors" quadrant. Their primary goal is to increase user count to get more ad impressions. All their content is catering around serving ads or showcasing sponsored content. This leads to popovers, modals, full screen takeovers and, my favorite, a "top 10 ways..." where you have click to the next "page" to get next item, thus loading new ads to serve. Some apps, such as New York Times or Ars Technica, will hybridize their product by convincing people to pay for an ad-free exerience or by offering some free content, then pushing remaining users behind a paywall. Although this form of revenue generation is at times annoying, and still slightly anti-user, it can be innocuous and provide a valuable way for content to be distributed which would otherwise not see the light of day.

You'll notice Facebook and Google operate between the yellow and the orange. That is because they are aggressively data mining their users in order to serve ads, not just display them. Facebook doesn't just say "hey, we have 1 billion users, let's just show them ads and make money that way". They are quantifying every user and offering a service to other businesses in order to craft targeted ads. Why is this considered harmful to users? The more data Facebook collects, the more accurate it's targeted ad platform becomes. This means that Facebook has a huge incentive to not only record and maintain a history of every social interaction of every user, but also to coax users into becoming more and more public with their sharing. The more public a Facebook user is, the more likely another pattern is registered by Facebook's algorithm which can be leveraged to serve that user a targeted ad. Collecting as much information as possible, not what is needed to execute a feature, is the prerogative, creating a ripe target for overreaching governments and hackers.

Now, let's imagine a Facebook that is not in the data mining quadrant, but is instead a hybrid of the top two: ads and direct pay. They could offer two versions of their product, free and $5 a month. If you pay, your data is completely protected and never used to serve ads. If you choose the free route, your anonymized data is used and stored to help serve targeted ads. With 1.06 billion users (as of this writing), if even 2% chose to pay, Facebook would generate over $100,600,000 per month.

Data mining encourages perpetual storage of user behavior and characteristics. It's not always nefarious, but if you build your business model around it, there is very little incentive to limit the amount of data you collect from a user; the more data you have, the more patterns you can ascertain. This is why Google moved away from offering just a basic search engine. They needed more avenues to collect user data.

Wait, aren't acquisitions supposed to be a good thing?

Acquisitions in the tech community are the holy grail. For founders and the few employees with decent equity packages, it is the big payout. But what does this mean for the users? When Instagram was acquired by Facebook, not much changed. But what about Sold, acquired by Dropbox? Shuttered. Dodgeball, acquired by Google in 2005, rotted away and the founders left to create Foursquare. Posterous, acquired by Twitter? Shut down. I Want Sandy? Dead.

Startups that focus on acquisitions over making money out the door treat their users like an asset to be sold, rather than improving the product for them. It makes sense: if your goal is to get acquired, your real customer is other companies willing to buy you. It's the short selling of the startup world. You create an idea, build up hype, get investors in, sell and smoke your cigars. Rinse and repeat.

This problem does not rest solely on startups, although they created the environment. The problem is that users are now trained to not pay for anything, and startups have a hard time even offering pay options. As a fun experiment, ask five of your friends if they'd pay for Facebook. I doubt you'll get a single "yes" out of any of them.

When a startup is created without an immediate business model, they are spending money they don't have from future acquirers or investment. They may even pretend to be user-centric in the beginning, but eventually they will focus all efforts on appeasing those that keep the lights on and the servers running.

The solution: pay for what you use

We need to reset ourselves and return to the normal way of doing business. A company offers a product I like, then I pay for it. The great news is that, for the most part, tech is cheap. You can get a lot out of a startup for very little, due to the low cost and high scalability of software.

If we turn around and start paying for these apps we love and use, we will see an increase in small, dedicated teams who love their products and keep improving them. We'll see features targeting users, not the advertisers or companies buying collated data sets. We'll see startups that are around for 5 or 10 years, not 6 months.

In a perfect world, the phrase "Don't worry, it's free" would arouse suspicion, not excitement.