Understanding the Cost of We Can’t Find Anything

One problem I often hear when talking with any organization about new solutions is understanding the cost and inefficiency of their existing way solutions, processes, or general way of doing things. In the past year or two I have used various general measurements around search to help focus the need for improvement not only on search, but the needed information and metadata needed to improve search.

We Can’t Find Anything

There is nothing more common that I hear from an organization about their intranet and internal information services than, “We can’t find anything.” (Some days I swear this is the mantra that must be intoned for an organization to become real.)

There are many reasons and potential solutions for improving the situation. Some of these involve improved search technologies, some improved search interfaces, or But, understanding the cost of this inefficiency is where I find it is valuable to start.

The first step after understanding you have this problem is to measure it, but most organizations don’t want to pay for that they are just looking for solutions (we all know how this turns out). The best method I find is walking through the broad understandings of the cost of inefficiencies.

The Numbers…

At Interop 2009 I presented “Next Generation Search: Social Bookmarking and Tagging”. This presentation started off with a look at the rough numbers behind the cost of search in the enterprise (see the first 16 slides). [I presented a similar presentation at the SharePoint Saturday DC event this past week, but evaluated SharePoint 2010’s new social tagging as the analysis focus.]

View more presentations from Thomas Vander Wal.

Most of the numbers come from Google white papers on search, which gets some of their numbers from an IDC white paper. I also have a white paper that was never published and is not public that has slightly more optimistic numbers, based on the percentage of time knowledge workers search (16% rather than the Google stated ~25% of a knowledge workers time is spent searching). There are a few Google white papers, but the Return on Information: adding to your ROI with Google Enterprise Search from 2009 is good (I do not endorse the Google Search Appliance, but am just using the numbers used to state the problem).

I focus on being optimistic and have I yet to run into an organization that claims to live up to the optimistic numbers or total cost of inefficiency.

We Know We Have a Costly Problem

This usually is enough to illustrate there is a problem and gap with spending time resolving. The first step is to set a baseline inside your organization. Examine search patterns, look at existing taxonomies (you have them and use them to some degree, yes?) and work to identify gaps, look at solutions like tagging (folksonomy) to validate the taxonomy and identify gaps (which also gives you the terms that will likely close that gap). But get a good understanding of what you have before you take steps. Also understand the easy solutions are never easy without solid understanding.

Evaluating what, if any taxonomy you have is essential. Understand who is driving the taxonomy development and up keep. Look at how to get what people in the organization are seeking in the words (terms) they use intend to find things (this is often far broader than any taxonomy provides).



12 responses to “Understanding the Cost of We Can’t Find Anything”

  1. todd@corvusconsulting.ca Avatar

    Strangely enough, I’m reading this just before listening to The Big Web Show’s episode on Content Strategy, which also happens to be a book I’m reading that makes content organization a full-on ‘other side’ of IA work; talk about synchronicity.

    Seeing the cost of poor findability scale up in large organizations reveals a bit of a dirty secret about the web: we think that because we usually find something sufficient with Google that search is working just fine. But peeking under the hood a bit as you do here reveals that it does need to be done better.

  2. edward.vielmetti@gmail.com Avatar

    Well of course we can’t find everything. That presumes some perfect world where everything in people’s brains is teleported into computers suitable for searching remotely.

    Have you ever thought about just asking a human being instead of doing all of this searching? The searches are failing because the people have an imperfect understanding of the systems. Someone else’s search might be better, and someone else might just know something with their experience and their brains instead of having to type something in and trust whatever comes back as truth.

  3. thomas@vanderwal.net Avatar

    I am not looking at the solutions (yet), but few organizations understand the cost today’s inefficiency is causing. This is the point. There is no perfect search, not even with asking people for help, which also have downsides, but can close the gap.

    Nearly every org I know complains that they can not find what they are sure exists. Few grasp the cost to that inefficiency. Fewer measure their own systems. While many want to improve on their known problems/symptoms, most balk at spending money on the solutions.

  4. ricardobeat@gmail.com Avatar
    ricardobeat@gmail.com

    Damn, why did Nielsen come up with this kind of productivity calculations? They are ridiculous. If those same employees spend 5 minutes a day picking their noses that means we have a $550.000 inneficiency cost? People are not machines.

  5. rlgschumann@gmail.com Avatar
    rlgschumann@gmail.com

    The maths don’t add up.

    16% of 40 hours is 6.4 hours, not 1.25… 6.4 hours per week amounts to 1.28 hours per day, and 20% of that (the failure rate) is just over 15.6 minutes per day being unproductive… or about 64 hours per year, based upon 50 working weeks per year.

    At US$30/hr (60,000 / 2000), this amounts to a cost of US$1920 per person per year lost to unproductive searches, significantly more than the US$375 you suggest… you therefore underestimate costs by a factor of approximately 5.

    Unless I’m missing something…

  6. brubrant@gmail.com Avatar

    @EdwardVielmetti you describe a perfect world as well…

    Imagine a huge company. I can’t just say: “Hey everybody! Stop what you are doing now and pay attention to me, I need to know where that document is! Anyone?”

    Of course you have social tools to do thinks like that without disturbing people’s work, but you need to be able to do that autonomously (trough search).

  7. thomas@vanderwal.net Avatar

    @ricardo This is about the tools people are using being the problem, not about the people. The tools help over what that state of things were prior, but there is an incredible amount of improvement that is needed. Measuring is how we improve the tools by identifying the gap between optimum and current state.

  8. thomas@vanderwal.net Avatar

    @Rob Deep thanks! I have been using these numbers, which were put together a couple years ago with a couple MBA and former consulting types as we were running through one of the white papers. I don’t have the original white paper (which didn’t extend the costs to per work per year.

    Oddly, not only have I not caught this, but nobody else has brought this up in the time I have been using this (about 2 years).

    I will be editing this post with the correct calculation. Bravo to you!

  9. rlgschumann@gmail.com Avatar
    rlgschumann@gmail.com

    The calculation is actually easier than I’ve broken down, though speaking in terms of hours and people is less abstract perhaps.

    With the figures you’ve been using the time ‘lost’ is 3.2% of total working time, so the cost is 3.2% of salary… US$1920.- in this case.

    20% of 16% of total = 3.2%

    Really surprised though that no-one spotted this beforehand though.

  10. duhl@rcn.com Avatar

    The IDC numbers on this topic come from Susan Feldman’s recurring research on “The Hidden Costs of Information Work” – Doc #217936 updated in May 2009.

    Her numbers for searching but not finding were 3.7 hours/week or $5974 per worker per year. It assumes a salary of $75,000 (including benefits) based on the Bureau of Labor Statistics’ annual salary and benefits figures for mid-2008 (www.bls.gov/data/#wages).

    Her research showed that on average people worked 46.45 hrs/week, with many working over 50 hours/week. Search ranged between 1 hour and 18 hours/day.

    Assuming an organization with 1000 knowledge workers — they’re losing nearly $6 million annually — searching and not finding. That is not a small number. Apply it to larger organizations and we’re talking tens to hundreds of millions of dollars in lost productivity.

    Furthermore, her research points out that since 2001 the amount of time people spend searching is rising, which is not surprising given that there are many more places to look — not just search engines, but social networks, blogs, daily feeds, etc.

    It would be nice to be optimistic about this problem… but the numbers indicate that it is far worse than what you’d like to admit. With people searching more, it will likely increase unless the tools become increasingly accurate or we become more proficient at using them.

  11. thomas@vanderwal.net Avatar

    I often lead with the optimistic, then point out it is much worse. I have had Susan’s past work and other doing similar research. Leading with worse case, nearly always leads to people undercutting the pain and as a result the whole. But, even optimistically the cost driven from inefficiency of search is incredibly large.

    When people seeing the problem go from understanding there is a problem to realizing their problem is much much worse than that, things usually start getting addressed. Organizations knew they have had problems, but really haven’t been moved enough to consider fixing it. Even with the optimal view presented here, the reality it has brought has been more than enough to get many moving and finally thinking about solutions and hiring people who actually grasp search well enough to deploy search tools that will make a difference.

  12. Jack Vinson Avatar
    Jack Vinson

    My news reader just popped up this article (possibly as a result of your rebuilding the blog).

    The struggle I have with these quantifications is that it makes the assumption that 100% "efficiency" of all workers is a good thing.

    A different question to ask to help quantify the problem is what can’t we do as a result of this problem? What is it that the knowledge workers can’t do? What does that take away from the business that people can’t find what they are looking for? Why were they looking for it in the first place? What is the real problem?

    The discussion of these kinds of questions can lead down some very different paths. Do we need a better job of educating people on information literacy (and storage and search)? Do we need to get people talking to each other? Do we need to redesign the workplace? Do we need a better search technology? Do we need better metadata? Does the stuff exist to begin with? Better display or visualization of results?

Leave a Reply

Your email address will not be published. Required fields are marked *