Why documentation analytics may mislead

By | 2011/12/07

I was rereading some material in the long-running do-people-read-the-manual debate (such as Tom Johnson’s If No One Reads the Manual, That’s Okay), and it struck me that there is an assumption that people on both sides of this debate are making which deserves some scrutiny. We all assume that technical documentation operates at first hand. That is, we assume that when a user wants help, they get that help directly by reading the manual or the help system or by watching a video, etc. I don’t think that assumption is correct. In fact, I’m convinced that it is naively and dangerously wrong, and that measurements and decisions based on this assumptions may be fundamentally flawed and harmful to a company and its customers.

Let me begin with a near-tautology: the influence of a work is not dependent on the number of people who read it, but on the influence it has. Here’s an example: it is one of the most influential books of the 20th century, arguable the most influential. You have either heard or read statements supporting or refuting its conclusions in the last week. Its influence has toppled governments and brought economies to riches and to ruin. You have never read it. Most of the people debating its principles, or acting on them, have never read it. I haven’t read it. You may never even have heard of it. This is it:

The General Theory of Employment, Interest and Money

The General Theory of Employment, Interest and Money

It is John Maynard Keynes’ The General Theory of Employment, Interest and Money and the degree of its influence, and the justification for my statement that you have heard or read an argument about its ideas in the last week, can be found here. The economic policies of the western world, and all the debates about them, are shaped by this old and, I am willing to bet, dull book that few people have read. Keynes book is almost incalculably influential today, and yet, virtually all of its influence is a second hand.

So, a work does not have to be widely read to be influential, and people do not have to have read the work to have their lives and their actions profoundly shaped by it. Any measurement of the importance or influence of a work that measures only its impact at first hand is badly flawed. As of this writing, John Maynard Keynes has a Klout score of 10. Mine is 32. Much as I would like to believe otherwise, I don’t think I am 3.2 times more influential in the world than the father of modern macroeconomics. Keynes needs to tweet more, clearly, but I can’t help thinking that Klout is simply measuring the wrong things.

Similarly, a user does not have to have read a manual (for manual, read any form or packaging of technical communication) for it to shape their actions and their behavior. I’m willing to bet that there is one person in your family who reads the manual for a new gadget and then teaches everyone else how to use it. (Since I am writing for technical writers, I am willing to bet that someone is you.) The manual reaches you at first hand, and the rest of your family at second hand. Your kids then probably show their friends how to use the gadget. Some of them will then pester their parents to buy the same gadget, and will then teach their siblings and parents how to use it. Let this process run for a while and you could well have thirty products sold and a hundred people knowing how to use them, all based on one person having read the manual.

Does this sort of thing happen with your content and your product? Almost certainly. What do you think is the first thing most people do when they need help? In all likelihood, they ask someone — a friend, a colleague. Isn’t that the way it works in your office? After that there is the use of social networks to ask people for help. Google’s Eric Schmidt was recently quoted as saying:

Social networking sites like Facebook and Twitter also allow users to leverage their social networks to find answers to their questions. Google is therefore competing with all methods available to access information on the Internet, not just other general search engines.

In The Tipping Point, Malcolm Gladwell describes the role of the maven in distributing information through a society. A maven is someone who is passionate about a subject, who learns and reads about it avidly and thrives on sharing their knowledge with other people. We have our mavens in technical communication — people like Scot Able, Sarah O’Keefe, and Tom Johnson who are constantly in touch with all that is being said and thought and written in the field. Want to keep up to date on all that is worth knowing in the technical communications space? Just subscribe to the twitter feeds of half a dozen or so of the top tech comm mavens. They read so you don’t have to.

Many people start, therefore, not with the manual, but with the maven. First there is the local office/neighborhood/bar maven. If that fails, the social web delivers mavens by mail. And if that fails, then there is still Google to search, where the answer that the user finds may or may not come directly from your docs (especially if they are behind a firewall), but may very well have come from someone who read your docs and posted about your product on line (that is, a maven). Much of the information that originates from your manual, then, may reach the user not through their direct reading of that manual, but via a network of mavens.

Information reaches users throught mavens.

Information from your content reaches users through mavens and social media as well as by direct reading.

In the little domestic drama described above, you are the maven, the one who reads and from whom knowledge flows to others. Every product has its mavens. (Why is there a 1-800 number on a bar of soap? Who would phone a 1-800 number for a bar of soap? As Gladwell explains, the answer is soap mavens.) Want to find the mavens for your product? Go on-line and find the forums where your product is discussed. You will find that there are a handful of people who answer almost all the questions. They are the acknowledged experts on the forum. People admire them and defer to their judgment. They are your mavens. They have read the manual.

Want to know what the best and most influential content in your documentation set is? Ask your mavens. If you need numbers, you will need to multiply your hit counts by the maven quotient. How do you calculate the maven quotient for any particular piece of content? I have no idea, but I would love to hear suggestions.


15 thoughts on “Why documentation analytics may mislead

  1. Pingback: Tech by juliakarolina - Pearltrees

  2. Scott Abel

    Great article. Good advice, too! This, from Scott Abel, who needs to get back to sifting through the data mines in search of content gold! Just wanted to stop by and say thanks for the mention.

    1. Mark Baker Post author

      Thanks Scott. And you are entirely welcome for the mention. It would be hard to think of a better example of a maven in this space.

  3. Tom Johnson

    This post is brilliant. Thanks for the insight. I hadn’t thought about how knowledge acquired in a second-hand manner can be just as influential if not more influential than knowledge acquired first hand. The problem is tracking this. Metrics govern funding, and if our only method for evaluating help is tracking hits, then low hits will result in low prioritization. I have no insight on how to measure second-hand knowledge acquisition, but I guess there would still be a correlation between help documentation and support calls. Thanks again for your blog.


    1. Mark Baker Post author

      Hi Tom. Yes, more influential, I suspect, in some cases. It comes back to what we discussed about a personal face. A maven may have got the information they pass on by reading the manual (though also by trial and error), but it is their personal authority that gives the people who consult them confidence in the information.

      Tracking it is, of course, a big problem. How do you measure a meme? It comes back to the problem of using proxy variables. There is nothing wrong with using proxy variables if you are confident that they are valid proxies; the danger is that we choose something not because it is what we should really be measuring, or a good proxy for what we should be measuring, but simply something that is easy to measure. Web analytics are about the easiest kind of measurement you can find, but if they are not a good proxy for the effectiveness with which an organization gets information to customers, then we shouldn’t rely too much on them.

      I think you are on the right track with your comment about support calls. It comes back to what you said in your post on rubrics, about measuring behaviors. There are other customer behaviors we could monitor as well, such as the questions they ask on forums, or the things they say on net promoter surveys. These things don’t produce such nice clear numbers as web analytics, and it can be hard to separate the influence of docs from other factors, but an imprecise measurement of the right thing is still better than a precise measurement of the wrong thing.

  4. pamela clark

    Thanks again for another great article Mark. You are so right about how knowledge spreads. Robert’s company gave him an iPad last December. Being an almost totally PC-oriented household before that, he found a lot of things about it mysterious and even frustrating as it didn’t fit his mental model of how it “should” work. The first thing he did was call one of friends who is a Mac “expert” of sorts. He claimed this was because he couldn’t even find a user’s guide, but I kind of doubt that he looked. He is a software engineer after all. My son tells us how to use features on our cell phones and is always sending us links. We found out about Dropbox from him too. Keep up the great blogs.

    1. Mark Baker Post author

      Thanks Pamela. My Nexus S phone and my Galaxy Tab both arrived with no documentation in the box other than the mandatory health and safety booklet, and a diagram showing how to plug in the charger. I rather expected that when I turned the Galaxy Tab on it would offer me a manual (as my Kindle did), but no. There seem to be more and more devices out there for which the manufacturers seem to expect that all information will come from mavens, and that the mavens will figure it out for themselves.

  5. Joe Gollner

    I think that you are right on target here Mark. There is quite a bit of research that falls directly in line with your argument – in areas of study like “innovation diffusion” and of course KM.

    And one part of your post struck home in particular – the one about there being a maven in pretty much every household and that it is probably the technical communicator. It struck home because this is partially true in our house. These is a maven for the equipment in the house and it’s not me. This probably proves the obvious point that I am not, in any way, a technical communicator. The maven is my son. When we purchased a brand new lawnmower, and a rather fancy one, it has a very specific starting procedure (for safety reasons). Not being able to guess them I was ready to return the lawnmower when my son emerged from the garage, manual in hand. “You are so pathetic! You didn’t even open the plastic bag that contained the manual, did you?” Guess not. I was not the maven but there definitely was one around.

    1. Mark Baker Post author

      Hi Joe. Thanks for the comment. You may not be a lawnmower maven, but you are, of course, a maven of intelligent content and content management. Sons, of course, are mavens of showing up dad.

      The book that originally got me thinking along these lines was The Social Life of Information. “For all information’s independence and extent, it is people, in their communities, organizations, and institutions, who ultimately decide what it all means and why it matters.”

  6. Tim Penner

    Well put, Mark. Although I think you’re giving the writer too much credit and the maven not enough. In fact, you make a fine case for documentation quality (what ever that is) being almost irrelevant.

    If the product is sound, then the doc need only contain hints and clues about its use because the maven will figure it out against all odds on behalf of his or her audience/network. Not being a mere conduit of our work, the maven is the product analyst with full credit for seeing product prospects and virtue, amplifying its value beyond anything a mere writer could aspire to by helping to define it’s social context. No matter how we fail to apprehend and satisfy our audience, the powerful maven will develop and sell the earthy truth.

    Sadly, I think you may have hit on a very inconvenient truth of our business.

    1. Mark Baker Post author

      Hi Tim,

      I deliberately avoided examining the full impact of the maven. I wanted just to talk about other channels by which information from a manual could reach the user without the user themselves reading the manual. But you are right that the maven may also generate and pass on information without reading the manual at all. A tech writer is, essentially, a hire maven, but, for some products at least, there is a community of mavens who may be far more knowledgeable and far more passionate about the product that the hired guns who write the manuals.

      Now that the general user community has easy access to these mavens through the net, it does raise questions about the marginal utility of hiring writers to produce official docs when the maven community will do the work for you, and perhaps do it better. My Nexus S and my Galaxy Tab both came without documentation other than a mandatory safety booklet and a single sheet on how to plug in the battery charger. I have yet to stumble upon any official docs for either of these devices.

      But I don’t think this can be the case for all products. I’m not sure the maven community for my dishwasher is as large or an active as the maven community for my smart phone.

      I’m planning to explore this further in an upcoming post, probably to be titled “Tech Comm’s Place in the Choir”.

  7. Pingback: Misleading Documentation Metrics | I'd Rather Be Writing

  8. Pingback: Making people aware of documentation, and encouraging them to use it | Communications from DMN

  9. Pingback: Differential Content Strategy - Every Page is Page One

  10. Pingback: Fewer People Read Longer Topics, and that’s Okay - Every Page is Page One

Leave a Reply