Why It’s Risky Business for Publishers to Build Their Own CMS

The following piece is a guest post. Read more about MediaShift guest posts here.

2014 was a crazy year for news technology. A new wave of publishers running custom software gained traction, while established news organizations continued to embrace open-source technologies. The overall awareness of the humble content management system (CMS) grew substantially and became an even bigger part of the “future of news” conversation. And building your own CMS is in style again, except now we sometimes call it a platform, especially in front of investors.

But building your own CMS is still very risky business. Technology is strategy, and culture eats strategy for breakfast.

How culture informs platform

Photo via Big Stock Photos.

I often feel like a news media tourist, since my consulting job takes me to many different kinds of organizations –– even in a single week. The most striking differences between organizations are in their internal cultures. The most successful companies I’ve known have stable cultures marked by patience, understanding and realistic expectations. Some have 20-year veterans on their digital teams! The least successful have impatient owners, have turned over most of their staff in the past year, and are seeking fast returns from new technologies. In the worst cases, where business and editorial are constantly battling, cultural deficiencies are bad enough that they’re visible on every article page. There’s often too many paid recirculation links, which make it seem like the organization is betting against its own editorial operation.

News organizations have three fundamental ways to deploy their websites. They can use an open-source platform like WordPress. They can build a custom platform on Ruby on Rails, Django or start totally from scratch by picking a programming language and opening a text editor. Or they can buy a commercial system from a vendor like Adobe, but that’s increasingly uncommon.

Then, from these starting points, there exists a very broad spectrum of customization. The point where the CMS ends and the custom work begins varies considerably from project to project. Plenty of large news organizations invest much more time and money in customizing existing platforms than smaller ones spend on creating brand-new ones. Deployed and customized correctly, any of these underlying systems could power any news site on the web.

But building your own CMS is especially hard. Even if you start from a framework like Django or Rails, there are significant under-the-hood development challenges to even the most basic CMS –– to say nothing of refining the user experience for reporters and editors. Yet the face value simplicity of storing and retrieving articles is a siren song to inexperienced developers, startup founders, and business leaders of news organizations. Even if you succeed in launching your own platform, you have still failed your business unless you can substantially outdo the open-source alternatives. What’s more, the open-source alternatives generally provide free and frequent updates. Frankly, you’ve also failed if you spend more to build your own platform than what it would cost to lease or buy a superior off-the-shelf product.

The perils of ‘roll your own’

Photo by Armando Torrealba and used here with Creative Commons license.

This is to say nothing of the most fatal flaw in the “roll your own CMS” plan. News organizations, especially ones with just one medium-sized digital property (this applies to many independent daily newspapers, for example), have a terrible time retaining technical talent. These organizations generally don’t employ more than a few such professionals, making the risk of attrition very high— and attrition is almost always cultural. Using a common framework like WordPress at least gives the next programmer a head start on understanding what his predecessor did.

Some news organizations take this challenge anyway because they want to open a new line of business and sell their platform to other publishers. One of my current projects started as a custom CMS by internal developers — a Django backend with an AngularJS frontend. It was meant to power a news website and also open an additional line of business for the parent company. But the developers working on the project quit with two months to go, so my clients lost their investment in that system and nearly had to forfeit their launch altogether.

In order to take your CMS to market, you need incredible tech team stability, market leadership, support capacity, and a great reputation. Perhaps the Washington Post, which may have these qualities, can pull off taking this route — but if they can’t, who can?

I would bet that Automattic, the company behind the open-source WordPress project, could become an effective publisher long before any news organization becomes an effective platform. They are certainly a platform — Automattic runs the massively popular WordPress.com self-publishing system, and recently purchased LongReads, too. And, through their WordPress.com VIP hosting program, they quietly run an astonishing number of high-traffic media sites — for example, TechCrunch, PandoDaily, Re/code and GigaOm all run on the same platform and have a surprisingly homogenous codebase. A publisher being the side effect of a platform seems much more likely than the opposite.

The push to be more than a publisher

Other news organizations, especially startups, often take the “custom platform pill” because they want to be perceived (and valued) as technology companies, as journalist and Gawker.com founding editor Elizabeth Spiers noted. Often this proposition is coupled with the idea that they will then sell their technology to other companies –– and sometimes it’s not. To succeed at building your own platform, you need plenty of cash. Big media companies have access to the kind of capital required to launch a successful custom platform, but startups largely do not. Fusion, the startup-like child of ABC and Univision, is a great example of this. ABC and Univision could bankroll a BuzzFeed-style custom web platform if they wanted to — and have certainly bankrolled a BuzzFeed-style stable of journalism talent — but they opted instead to run WordPress. Likewise, Atlantic Media’s Quartz — which broke ground on reader experience and is often imitated –– picked WordPress. (Automattic hosts both of these sites, too).

For startup news organizations, the primary way to get the money to build your own platform is venture capital, which is rightly skeptical of news startups regardless of their technology. This isn’t to say that a new news organization can’t succeed — it’s just unlikely to succeed at venture scale. The startups that get funded largely have compelling technology stories. Would BuzzFeed have raised $96.3 million on WordPress? Probably not. In fact, a great concept for a news organization without a good technology story would have a very hard time raising money at all.

Any news organization, from startups to conglomerates and everyone in between, will inevitably plant their expectations for themselves into their choice of platform, so the decision must be undertaken with clear eyes about what can and can’t be accomplished. If your mission is outlier-or-bust and you need venture money to get there, do what you must. But otherwise, think long and hard before turning down that road — your culture is far more important than your technology, and you could put it at great risk by striking out on your own.

Austin Smith is Managing Partner and Co-Founder of Alley Interactive, a digital agency that designs and builds large-scale content sites for top digital publishers and leaders in news media. Follow Austin @netaustin.