Comment by adrianb | Hacker News Reader

Comment by adrianb | Hacker News Reader

Parent Full thread

The old software development model where users are buying the license for a specific version of a program and then use it and only come back to spend more when they need an upgrade is pretty much dead now. Most software moved to either free downloads with paid support (even Windows these days?) or to membership-based model (Adobe, Office365).

The reason is that as you add features and fix bugs in a software, fewer users need the upgrade (why spend another 150$ on this year's release for some features I never use?). But the development team still needs to be paid and the advanced features they add each year tend to be harder to implement, so over time the costs are increasing while sales go down.

Adobe might be successful with this business model because they have a strong market in professionals but they definitely lost the amateurs and hobbyists.

EDIT to add: also it didn't help that CPUs and OS APIs haven't had major changes in more than a decade. There was no major reason for a large majority of users to upgrade (such as the move to 64 bit CPUs).

> But the development team still needs to be paid and the advanced features they add each year tend to be harder to implement, so over time the costs are increasing while sales go down.

Do they need to? I mean, if users don’t see the value in spending $150 for a few new features they don’t use, that says that maybe the ROI isn’t there for those features.

Think about the marginal value of each hour of developer time.

Value = number of feature’s users across the product’s lifetime * utility of the feature per user / cost to implement that feature.

Early in a product’s lifecycle the features added have high utility (you add the most important features first) and most of your users haven’t bought your product yet. (So features are a long term investment). Marginal expected value is high - assuming you get the sales. Later in a product’s lifecycle the features are more marginal (and often more costly to implement). If users don’t want to pay for the upgrade, what they’re saying is the new features being added aren’t worth the development cost. If the user base is still growing it might still be worth adding those features. But at some point you should entertain the idea that the product is essentially done. The best play might be to move most of the engineers to a new product. Keep a small team which can fix bugs, renew google maps API keys and add ARM support and so on and call it a day. Profit should go up at that point because you should keep getting sales while your costs plummet. Over time sales drop as your old product has sold to most users who want it, and in that time the income can be used to bring a new product to market.

This is the model the video game industry uses (to great effect). Why are we so allergic to the idea of “feature complete” in software?

The alternative is what Adobe does - try to wring ever more money from your customers while providing them less value each year. It’s no wonder customers are desperate for alternatives to their products.

> (even Windows these days?)

Windows 10 Pro Retail is still $200.

However it was quite easy (still is?) to upgrade a Windows 7 Home OEM to a Windows 10 Pro... for free !

Reply on news.ycombinator.com