Software Business Longevity

Reading about Artifact’s announcement of their imminent shutdown was a bummer.

Rereading some of the media coverage of the app when it came out of beta a year ago, much of what the founders said at the time made sense. Social networks, particularly when there are already a handful of entrenched, mature services, need to bootstrap social connectivity via utility. Artifact started by constructing a categorization of news articles and media outlets algorithmically—a Tiktok-esque spin on a news feed—and eventually added comments and user link submissions as social inputs.

Artifact became one of my daily news apps in the past year; its discovery mechanisms came at a good time when Twitter’s role in news curation waned under new ownership. Yet, the evolution and progression of its community reminded me a bit of Quora’s downward arc of quality1: it’s hard to maintain a high bar at scale, and inevitably a larger user base will find ways to game the system faster than its developers can tweak it.

You’d think that the folks who built Instagram, who garnered high enough growth rates to scare Facebook into an acquisition, could repeat that success on demand. The founders were wealthy enough that they self-funded the app from inception, so the shutdown here is less about external investor pressure to accelerate growth rates, and more about deeming that the opportunity costs for the team were not worth their while. In an alternate universe, Artifact would have been an experimental app, from a startup “lab,” that gains enough traction to spin out an independent team but ultimately fails to reach escape velocity.

It used to be that only tech nerds cared about whether a service was sustainable via its business model. That’s still largely the case, though with apps having come and gone in the past decade, regular users now also have ample experience with favorite services discontinuing or degrading gradually. Most of these services are cheap or free, so the costs aren’t so much monetary, as they are time and attention and data. If the premise, like in Artifact’s case, is to personalize from revealed preferences on user actions, those signals will disappear with the server instances.

There are 2 main ways to hedge against this possibility: own the app or own the data. In the former case, that usually means looking for an open-source alternative and hosting it yourself. WordPress and Ghost are great blogging/CMS systems in this regard, and their commitment to keeping their product as an OSS project is one of the reasons that I’ve continued to lean on WP for this blog. With the latter, apps that use—or support export of—simple, standard file formats work well, particularly if they rely on common cloud services for multi-device syncing. Obsidian does this well, and its neutrality in data is one of its selling points.

But it’s preferable to have a great app or service just stick around. Apps like Omnivore and Raycast know that their users are wary of free, so they make it a point to address the issue upfront, especially during the honeymoon initial phases when personal passion and venture capitalist excitement seed the 1.0.

Hmm, maybe there’s something to finished software after all.


  1. I wrote that piece over a decade ago and haven’t paid much to the site since; it sounds like it’s even worse now.

Share this article
Shareable URL
Prev Post

Batched Reading Efficiencies

Next Post

Site Under Construction

Read next