The Need to Weed

cleaningOne of the conversations that stayed with me from a eDemocracy camp gathering several years back was the clear disconnect between the push toward transparency – often read as put it all online – and the struggles against info overload.

From both the web consumer and web manager point of view there is simply a huge problem with the fact that too often online information isn’t relevant or easily found. By relevant, I mean the right information at the right time – is it meaningful content. Easily found means that there is a strong scent for the information that allows you to easily navigate from the home page to the right landing page.


But what happens when there are content problems? Content can be irrelevant. It can be out of date. It can be redundant. Or it can be inaccurate.

The driving force behind efforts at the US Air Force to consolidate and improve its Intranet a few years back was a pair of insightful reports cataloging woeful content problems. At one Major Command a KPMG study found 40 percent of its content was irrelevant, out-of-date or redundant. In a separate study the USAF Inspector General estimated 70 percent of the content on the public sites had the same issues. These wake-up call reports helped rally the funding and leadership necessary for change.

When the online consumer is lead astray by inaccurate content it can have high-risk consequences. Take an example such as the technical manuals at the US Air Force. When old manuals are published online and used by a technical officers in charge of repairing planes, a critical repair could be missed or handled wrong. That means the plane could crash and perhaps kills its pilot.

Or on a more mundane note, you look up instructions to trouble-shoot your new software program and you find a page with last year’s version instead of the up to date one that has an easier and quicker solution to your problem.


As we ponder this month’s Earth Day celebration, it is worth taking the time to consider “weeding the garden” – a phrase that has become popular with those who manage Web sites.

As Gerry McGovern reminded a recent Webinar audience in talking about the need to weed, when you let everyone in your organization have space on your site, you get complexity not simplicity. Instead web leaders should respect consumer power over organization power. Strip away the froth and non-essentials and promote the essentials: links that let you do the top tasks on that site.

McGovern referenced a study on the paradox of choice. When there are too many things to choose from the consumer tends to freeze up and moving ahead becomes more difficult. A grocery store had a taste test on jams. When they put out six different jars, they found 30 percent who stopped to taste purchased some jam. But when they put out 20 jars to try, only three percent made a purchase.


At the Webinar, a case study from Microsoft Office was a featured. After they become alarmed at analytics reports which showed dissatisfaction was on the rise, they realized it was time for a change. At the core was the problem of a culture geared toward the classical satisfaction metric – instead of a different metric to look at what makes people dissatisfied. Case Study: Microsoft Office

At the Microsoft Office web team, their vision was, “Give users what they need”. So content publishers heard, “publish lots of cool content – more, better, faster”. Page views were up, but satisfaction was flat. Why? Customers were overwhelmed and couldn’t find stuff, or looked at the site and concluded it wasn’t what they wanted. The first attempt to fix this was also a classic web team reaction. Content was revised. Keywords were added to search. Links were added in a redesign. The result: satisfaction continued to be flat and the site was even harder to keep fresh and relevant.

So they concluded it was time for a radically different approach. They decided to recognize, measure, report and share Dissatisfaction as their metric. To increase satisfaction the edict was publish more pages. To LOWER dissatisfaction, the edict was to REMOVE bad pages or reduce their use. What’s more, they got a mandate to weed immediately, and publish only as soon as possible.

It was a revolutionary change in culture. The reaction was alarm. The easy thing was to leave things in place on the theory that there were “good pages” that some users would want to find. But positioned as an experiment, these Web leaders forged ahead with a pilot and aggressively began removing content and taking action to cut views.

They created a list of pages that provoked high dissatisfaction.  The team’s new mission provided options.  Option One: remove these pages without making any revisions. Alternatively, the second option was to cut the views by hiding it from search by taking action to re-title the article and remove the links. There was NO OPTION available to preserve the status-quo and take no action.


Before the experiment, their metrics showed both page views and dissatisfaction on the upswing. Six months after they weeded the garden, overall dissatisfactory page views dropped to a mere three percent from a level of ten percent six months before the weeding project.

On specific experiments, the results were even more dramatic. On the function topics area in Excel, the Microsoft Office team found satisfaction rose 423 percent and dissatisfaction dropped 42 percent.

The lessons drawn from this experiment included

  • You cannot succeed with satisfaction metrics alone. You need to check for weeding all the time – it is as important if not more important than publishing.
  • Cross-functional collaboration is a must – Putting steam behind a movement to weed takes executive leadership and cooperation
  • You gain productivity with fewer pages to manage – showcase the benefits to realign the incentive structure that usually is built entirely around publishing, not weeding
  • Users do not miss the de-linked or removed pages  – so fear not. Change will not bring a revolt from the outside

What Microsoft’s case study confirmed to me is that it is time for web managers to consider a new approach to managing their properties. What’s more, to blindly embrace the goals of the transparency movement without considering the very real problems of information overload is a problem that cannot be easily brushed aside.


Where do you start? When you can describe the problem you own the solution.

Here’s the problem: of the 24,000 federal web sites that exist today, how much of the content is irrelevant, out-of-date, redundant or inaccurate? Even a estimate that is rough would help rally the web community. I’d like to see someone in the federal web community steal a page from the USAF playbook and try to scope the problem.

A web audit to spot-check the issue could be the just the report needed to spark action. Whether you are a public or a private site, I’ve found that an executive wake-up call is often necessary to document the scope of the problem and put a bulls-eye on it.

First, you must understand the answer to this question: when your customers arrive at your web site, what are their top tasks? If you want to create customer-centric, task focused Web sites that generate results you’ll be proud of, it’s time to find out.


When you structure a site around top tasks, and put a premium on simplicity you prove you’re your allegiance is with customer power, not organization power. If content is redundant, out-of-date, irrelevant or inaccurate it is time to let it go.

The need to weed the garden for the collective Web sites published by the public and private sector has never been greater. So get out your spades and shovels and let’s begin.

Do you think the transparency movement pays enough attention information overload and the need to weed?