Book Review Part II: Content Strategy for the Web–Minimilism
One of Krisitina Halvorson’s first points in her book Content Strategy for the Web is that it’s cheaper, more relevant, and easier to produce the minimum amount of content on a Web site. I can’t agree more that minimalism is cheaper and tends to produce more relevant experiences (Kristina uses the term user-friendly, which makes me cringe, so I use more relevant.) But I think it’s actually a lot harder to put systems and processes in place to keep content inventory to a minimum. I’ll talk about these points after the break.
Minimalism is cheaper
Keeping content to a minimum is obviously cheaper. Content strategists, writers, editors, search and social media experts, architects, designers, production people, content management people, Web analytics experts, webmasters and their associated project managers are not cheap.
Hosting is not cheap either. If you saw how much IBM pays for content hosting every year, your jaw would drop to the floor. When you add in applications such as search and personalization, content becomes very expensive. It’s cheaper than print, but only by the cost of paper and ink, which is a tiny fraction of the cost of content. If you can reduce your content inventory by a significant amount, you can save a lot of money (and energy–data centers can be energy hogs).
Minimalism results in better relevance
Minimalism tends to result in better relevance because needless duplication is one of the most difficult things for users to deal with. Take search, for example. If you have five pieces of content (or URLs for that matter) that all focus on the same keywords, the link juice will be spread out among the five. Thus, none of the five will rank highly enough in Google to be findable.If you can focus your architecture around one or two pieces of content, you can concentrate the link juice to those URLs and improve Google results, and findability.
On your internal search engine, if the user can’t tell the difference between different pieces of content in the SERP, you will have a low search satisfaction for your internal search engine. Forcing the user to click a result to determine relevance is a bad user experience. But, in our experience, apparently duplicate content does just that. Note, it doesn’t have to be duplicate, it just has to appear to be duplicate content from the users’ perspective.
And those are just on the search side. Duplicate content is just as much friction for navigation as it is for search. In our book, we make a point of saying that if a user comes across the same piece of content twice, the second time it will be less relevant, assuming it was relevant the first time. You don’t consume the same content twice; there just is no time, unless it’s a checklist or something. In almost all cases, apparently duplicate content reduces relevance.
Minimalism is harder
When I started at IBM in 2004, we had some 12 million pieces of content on ibm.com. I didn’t see the number of URLs. But based on the ratio we have now, we had something like 8 million URLs on ibm.com. The other 4 million pieces of content consisted of white papers, demos, presentations, data sheets and other types of collateral. At the time, we had some 30,000 content owners. (Not that content was their full-time job. It might only be 5% of their job.) I’m not counting our intranet, which is another boatload of content.
I am proud to say we have cut our ibm.com content inventory in half in the time I’ve worked at IBM. That has happened while the company has acquired several large software and services brands and brought their content into ibm.com. Acquisitions and a more federated content model have increased the number of content owners to something like 45,000 (not counting the thousands of intranet content owners).
Most of the inventory reduction was not my doing. Smart executives set aggressive schedules to reduce content duplication and to archive content that is either out of date or no longer gets traffic.They did it for the reasons Kristina sites: Content minimalism is cheaper and tends to produce more relevant content experiences.
In 2009, I was assigned a project to help ensure a consistent content expiration process, ultimately to reduce our content inventory. Our audit revealed that most of the work was already done. But the audit highlighted the challenge involved in content minimalism. Every brand and business unit had its own process for setting time limits for content to be live, and for notifying content owners that content was about to expire or had achieved the dreaded zero traffic threshold. They all had processes for forcing expiration if content owners didn’t respond. Like most IT companies, IBM encourages a lot of movement from division to division. So something like 30 percent of owners listed in the source code for pages had changed jobs by the time the notices started. Tracking down the owner is a onerous process. Getting them to admit that their content was no longer valuable is even harder.
I’m happy to say that we now have one content expiration process. And it works pretty well. But we still cannot get much leaner than 4 million URLs and some 6 million pieces of content. Sure, we have some apparently duplicate content. But in those cases, it serves a different purpose or audience. So it is necessary. I’m sure we could reduce it further if we hired a team of auditors and architects to determine when and where content can be reused for multiple purposes or audiences. This is harder than you might think. For reasons I will explain in some future post, it would require the kind of technology that hasn’t been released by even the most sophisticated search companies.
Considering that IBM produces thousands of pieces of content per week for ibm.com (not to mention the intranet), just holding steady on our content inventory is a win at this point. Yet, when I told Kristina how much content IBM has in its inventory over lunch last year, she looked up at me through her eyeborws and said, “And this is a good thing?”
I just don’t think she had this kind of massive site in mind when she wrote her book. Even for smaller sites, it is much easier not to build all kinds of auditing, archiving and expiration processes into the content management system. That’s essentially why IBM had so much content when I started. They hadn’t yet to put those processes in place. One reason, I suppose, was that developing and implementing minimalist processes is expensive. They pay for themselves in reduced content resources and hosting costs. Most importantly, they more than pay for themselves in increased revenue: More relevant content makes companies more money. Until we netted out the business case, it was just easier to leave the content hose on full blast.
IBM is proof that content minimalism is worth the investment. I just don’t want people to think it’s easy. It’s hard. The more content owners you have, the harder it is.Validating that we had the makings of a good process consumed more than half my job for a whole quarter. I’m sure I’ll need a similar effort in the near future to refine our process and to make it more intelligent. Like all things on the Web, it’s not a one-and-done, set-it-and-forget-it process. It’s iterative. That makes it all the more challenging.