Web task management principles
Web task management is about managing your website around common tasks. Success is measured on the completion of these tasks.
What is different about web task management? Traditional website management focuses on managing the technology and/or the content. Such website management approaches are generally project-based.
Under traditional web management models, for example, launching a search engine for the website is a project. Once that search engine is launched, nobody is made responsible and there are no quality measures for success. The search engine is simply left there.
Equally, there are various projects concerned with putting up content on websites. The focus is on getting the content up as quickly and simply as possible. This sort of approach is particularly found when new content management software is being installed. Transferring lots of content to the website becomes a big project.
These management approaches fail because they manage and measure the wrong things. If you manage purely from a technology point of view, then the technology itself becomes the focus.
Organizations have often bought overly complicated content management software because of the belief that if you buy the ‘right’ software, you solve the problem. Only passing attention was given to what customers wanted to do on the website. The tool itself became the focus.
It is an equally bad idea to manage from a content point of view. Communicators love to communicate. An intranet or public website can seem like a nirvana for the avid communicator. Vast quantities of content get published, not because of a clearly defined need, but from a ‘have gigabytes must fill’ mentality.
The fatal flaw in managing from a technology or a content perspective is the management metrics these approaches deliver. If you manage from a technology perspective, then the metrics are nearly always volume-based. It’s about the number of documents that are published, or the number of searches that are carried out.
Managing from a content perspective is even more volume-based. When I hear many senior managers talk about their websites, I am surprised that they are still quoting the utterly useless measure, HITS. (HITS stands for “How Idiots Track Success.”)
Why is this? Well, many years back, probably at the beginning of the Dot Com boom, there was a request from the CEO’s office for something to say about the website at some conference. The web team pored over the website log data and saw this wonderful metric called HITS. And why was HITS so wonderful? Because it was a very big number.
And so began an obsession with volume. The more volume the more success. HITS may have taken a back seat in many web metric models today, but page impressions/views, as well as repeat and unique visitors, are certainly thrown around with gusto.
This approach to measuring a website’s success will surely end in tears. For a mature website, measuring the increase in page impressions is as likely to reflect the failure, not success, of the website.
Jacinta Thomler says:
Added on May 20th, 2007 at 10:59 pmsure, measuring page impressions only reflects volumes, not success. But CEOs still want a easy figure.
What is a realistic alternative?
Stephen Gosden says:
Added on May 21st, 2007 at 8:10 amJacinta beat me to it - that was my question too! If not hits then what? How do we measure progress or assess whether or not our web sites are delivering what the customer needs or wants?
And how do we assess whether or not we are increasing our reach? Is “unique visitors” a good measure?
Vivienne Maguire says:
Added on May 21st, 2007 at 8:17 amI agree entirely, any direct response medium should be measured on acquisition/conversion/sale whatever you like to call it. There are many factors which are hampering the growth of a strong online market - this is definately one of them.
Michael from UK says:
Added on May 21st, 2007 at 9:52 amAs usual, great atuff from Gerry.
Realistic alternatives to visitor metrics? I would suggest looking at evidence of task initiation/completion from the user. Sales, obviously, and in the public/not for profit sector that translates as eg. sign ups to email newsletters, booking requests, participation in consultation, or use of fault reporting tools (eg. potholes, faulty streetlights etc).
In other words you are looking for real activity or commitment on the part of the user, beyond mere passive viewing of pages.
Of course, you haver to have those choices available in the first place.
I’ll be interested to see if I am thinking along the right lines.
Dick Nieuwenhuis says:
Added on May 21st, 2007 at 12:09 pmThe alternative is a combination of varies types of statistics (unique visitors, geographical spread, for example) and trends. And some form of user satisfaction surveys (mail or other).
But statistics are what they are: lies, big lies and statistics.
Claudia Allen says:
Added on May 21st, 2007 at 1:41 pmPages that rank high in Google; click-throughs on the nav bar or in links within copy tell us people are looking at/using the web pages.
Karen R says:
Added on May 21st, 2007 at 1:55 pmSo it’s not size that matters, it’s how they use it, eh? ;^)
On our government site, although we’re working at identifying top tasks that involve processes (signing up for our news releases, submitting a FOIA request), I suspect many of our visitors’ top tasks fall under the nebulous heading of “finding information.” That’s rotten hard to quantify.
In the meantime, I’m reporting traffic trends, top search phrases, and most popular pages. We do use a customer satisfaction survey, but our agency’s site is so large that I never see results relevant to my corner of it.
I am learning to embrace uncertainty.
Gerry McGovern (blog author) says:
Added on May 21st, 2007 at 3:19 pmOf course, if bulks measures are not good, then what? Task completion. But by task completion, I mean something broader than is commonly understood.
Michael is definitely on the right path here. He mentions reporting potholes. Well, if you were a local government website, you could ask a sample of citizens the following: Try using our website to report a pothole that needs repairing?
I think at a basic level we need to define the really important tasks of the website. (Again, these don’t have to be transactional.) Then, test to see can our customers complete these tasks.
And the only way I know how to do this well is to get a sample of people to actually attemtp these tasks. Sure, this is a cost, but otherwise we are blind to whether the website is working at the most basic level.
Karen mentions top searches, which can be useful. But if you see something seached for a lot wouldn’t that indicate that there might be a problem in the website design. And if you solve that problem, wouldn’t you expect that search behavior to decline?
Marc Bernstein says:
Added on May 21st, 2007 at 4:12 pmI agree that transactional measures are an important component of website usage. As a not-for-profit, member-based organziation, we report on hits, but should be supplimenting with trends in transactions. For example: percentage of customers joining/renewing online, numbers of customers who update their contact info online, people subscribing/unsubscribing to newsletter via the website, etc. We have these metrics, but organizationally don’t report on them consistantly.
I’m not so sure that high volumen search terms means that there is a design problem, per se, but gives insight into what people are looking for. Many people use search as their primary navigational method, not bothering to look through traditional navigation. Recurring search terms could give insight into member needs which are opportunities to market and develop product offerings, member outreach and communication. These could drive more traffic to the site for that topic, resulting in more search activity on the topic, not less.
Another useful search metric is how many times the user clicks through any of the search results.
I agree with Gerry that ultimately, you need to be conducting some in-person marketing research/usability studies to really understand what people need to do on your site and then watch them do it, either easily or with difficulty. This is costly, but the gains to be made are potentially large.
Gerry McGovern (blog author) says:
Added on May 21st, 2007 at 4:46 pmMarc is right in his comments about search. A high volume search for a particular word does not necessarily mean that there is a design problem.
We do need to analyze search, though, to understand what is driving the high volume searches.
Dick Nieuwenhuis says:
Added on May 22nd, 2007 at 3:28 pmAnd there remains a problem in as far all surveys and stats tell you something about those that DO come to your website. But what about the other ones that you would like to receive on your pages? How do you know about that??
Brian Anderson says:
Added on May 23rd, 2007 at 6:00 amI love problem solving so here’s my two bits ….
One of the biggest problems on the Web for site visitors is lack of direct evidence that their actions have made any difference or that anyone is listening. Case in point, could the people that reported potholes receive a thank you note, online, and a published list, on the gov’t website, of these identified potholes and expected work dates? And maybe who it was reported by? People love to see their name in print or even on a white screen. It signifies that they exist and that they have been heard. Everyone likes that. The objective: reward by recognition.
Setting out the important tasks for a website (yes, kudos to Gerry) provides focus and a basis for the conversations to come. In this scenario, each task is a conversation with a site visitor.
In sales, which is my background, we spoke of ‘earning the right’ to meet with a prospective customer. Getting into see them wasn’t a given. To achieve that task, a salesperson has to convince the prospect that they had something of value to discuss. In websites, earning the right might equate to demonstrating that you genuinely care about their experience while on your website.
Jumping ahead to measuring the quantity of sign-ups for surveys or newletters may be expecting too much too soon. You haven’t established enough trust yet.
For example, after searches are conducted, maybe a box pops up with an inquiry asking people to indicate if they ‘found what they were looking for and to choose the problem from a list of choices or to write a sentence or two in a small 2-line box.’ In this way, the task request is small, and, if worded appropriately, communicates your concern for their time, patience, and website experience. This is a sales technique by the way; it’s called ‘nibbling.’ In this context, you encourage the “client” to complete a small task on the way to larger ones - but you can’t rush it. In our web context, since it is all about the customer, this task will be helpful to them. You then build upon that with other customer-centric scenarios to larger tasks.
Communicating back to clients, via the site, that their recommendations have led to this or that change may sway more people to get involved. Thus, their actions are rewarded and everyone is happy.
In some ways, this is about getting more personal over an impersonal communications medium and inspiring more trust by initiating more feedback from the site itself.
Well, I am most interested in your comments.
Gerry McGovern (blog author) says:
Added on May 23rd, 2007 at 6:42 amDick identifies a basic flaw in all web stats. They only tell you about the people who do come to your website–not the people who have stopped coming, and not the people who should come.
Also, stats are so flat. I’ve looked at lots of stats overt the years and found them very hard to understand, particularly where they deal with path analsysis. “They clicked here and then they clicked here.” What does that actually mean? Did they get what they wanted? Were they frustrated?
Brian raises some very interesting ideas. What does task completion actually mean? That you reported the pothole on the website? Or that the pothole actually gets fixed? And it would certainly be very useful to get feedback that the effort you went to to report the pothole was worth it.
Brian also suggests finding ways to have small interactions with the customer online to judge how well they are doing. I think this approach certainly has promise.
It all brings to mind a key point for me. We need to be constantly thinking of the customers’ experience, rather than our need to manage the technology and the content.
George Thompson says:
Added on June 29th, 2007 at 2:10 pmWhat is your point? What are the metrics for measuring in a task-based environment?
Gerry McGovern (blog author) says:
Added on July 1st, 2007 at 2:34 pmGeorge, the key metric of tas management is task completion. You give a sample of your customers common tasks and you measure whether they can complete them and how long it takes.
Craig Thomler says:
Added on May 23rd, 2008 at 8:04 amAfter more than 12 years in the area, I’ve concluded that there’s no single metric that will provide all the answers for web measurement.
My approach is to start by defining the goals of the site - is it designed to achieve sales, behaviour change, inform people, build networks or simply to make people feel happier (just some of the potential goals).
Once the goals are defined you can at least test the available quantitative and qualitative measures for relevance and select the measures that most accurately tell you if you’re getting your ROI.
To give a few examples, I oversee both an intranet and an information-based website (with some transactional components).
For the intranet, we track user satisfaction through regular qualitative benchmarking. How happy our staff are with the system is a key indicator for us of whether the intranet is performing as if they are not happy they’ll use it less, therefore reducing the effectiveness of the intranet as a comms and informational distributional tool.
We also measure tool use and task completion - particularly use of address book, calculators and forms and use of process documentation that our business expects staff to rely on.
Finally we put a high measure on awareness building and informational gaps. This is tracked through a variety of means - search reports (what are people looking for, what are they NOT finding), news headline click-throughs and the like.
Now senior management is most interested in the number of visits, the duration of those visits and the top pages/tools used - so we provide those figures, but there’s a lot of detailed work under the hood which supports user access growth.
For our website we track actions and reactions. Customers don’t just find our website, in many cases they are lead there through various comms sources and are looking for specific information or activities. I can track the effectiveness of our media and marketing through measuring responses in our website, changes in search behaviour, variations in tool use, etc.
Speaking of tool use, I also work closely on the measurement of a secure transaction-specific site. In this case what is particularly important to us are returning customers, task completion and, again, gap analysis (what tasks are people wanting to complete). We also look closely at the workflow funnel (for a commercial site this would be the sales funnel) to analyse which steps in the workflow cause the greatest drop-out rate so we can pinpoint the steps/screens that require the greatest level of rethinking or contextual help.
Sign-ups are significantly less important, though still relevant for awareness and attraction. However it’s no use signing millions of customers if they do not use the site on a semi-regular or regular basis!
A few thoughts I always keep in mind around web management are;
- The issue is not finding relevant measures, but selecting from the many different measures available to build the most accurate picture of what is really occurring.
- Trends are more important than pinpoint data, just as cashflows can tell you more about an organisation than a single balance sheet.
- You cannot manage what you do not measure - and you must measure the right things to manage well.
Cheers,
Craig