Several of our engagements and serious consulting proposals recently have been very much focused on helping our clients build out a measurement department. Obviously, this type of work is bittersweet (or perhaps I should coin an oxymoron and say "richpoor") for us. On the one hand, the work is interesting and useful. It’s even remunerative enough while it lasts. But the while it lasts is the obvious sticking point. It’s like being a veteran quarterback brought into a team that just used a #1 draft pick on a new young quarterback. They pay you well enough – but it’s only a matter of time before you hit the bench!
As part of that work, of course, we’re sometimes blessed with the opportunity to advise a client on how a measurement department should be organized, what kind of people it should hire and where in the organization it should fit. I thought I’d take up a few of these topics in the blog and I’m going to start with some thoughts on where a measurement department should fit within an organization.
Of course there isn’t one right answer to this question. And I really mean that - I'm not just saying it to provide the obligatory wishy-washy qualifier. I’m going to talk about where, in general, I think it might make sense to put a web measurement department and I’m necessarily going to make some very broad generalizations about IT Departments, Marketing Departments, Senior Executives, etc. But since all these organizations are made up of people – and people are always different – my broad generalizations won’t hold in lots and lots of cases.
As an example, we work for one company (and a public company at that) where one very Senior Executive is so smart, powerful and hands-on that the entire web measurement effort is pretty much focused (rightly, I think) on providing answers to his questions. That isn’t a model I’d necessarily recommend. But it’s the right model for that company with that person and it insures that the web measurement is valuable and directive for the company as a whole.
Where to put web measurement isn't really the first question you should ask. Before you even start thinking about where you’re going to place a web measurement department, you need to think carefully about whether you should have one. And I don’t mean whether you should outsource to someone like us (that’s an entirely different question) – I mean should web measurement expertise be spread out through an organization – with Business Units or Managers hiring it where and when they feel a need.
I’ll call this a de-centralized model. In a de-centralized model, you don’t have a measurement department, you have pockets of expertise – presumably where they are most needed. There are some real advantages to de-centralization. I’ve never been a big believer in Matrix solutions – and the de-centralized model insures that the web analyst will be directly responsible to the information consumer. That’s a very good thing. It means the analyst is closer to the ground and much more likely to understand the business issues because they’re his boss’ issues. It also means that each BU will only keep web analytics resources if they seem to be adding value. And no one should have a better idea if that’s true than person whose supposed to be using the information. Because web analytics is all about knowing the business, keeping practitioners responsible to the business owners is bound to be advantageous.
There are a few things you lose if you go to the de-centralized model. First, it makes it much harder to build expertise. You may have one good analyst – but if that analyst is siloed in a BU, no other BU can take advantage of that expertise to train new hires. In addition, having web analytics expertise at the BU level can make it hard to standardize approaches and metrics across the organization. I was talking to someone at the SEM Pharma conference (I was speaking on the Seven Deadly Sins of SEM Measurement - drop me a line if you'd like a copy of the Powerpoint) whose organization has upwards of five hundred web sites! In that kind of environment, how could senior management possibly insure the integrity of the numbers without some kind of central analytics organization? Finally, a de-centralized model may make it hard for smaller BU’s to meet any web analytic needs. Large BU’s may have enough demand for an FTE – but if they simply fill their slots then smaller units will be left out in the cold.
Of course, a very large organization could also adopt a hybrid model. Large BU’s would have de-centralized resources directly under their control. These would handle the analysis for their units. There would also be a corporate unit for insuring standardization and providing shared resources to smaller BU’s and training across BU’s. Naturally, your central organization will always be fighting with the BU’s about tools, standards, etc. But that doesn’t make this a bad approach. Indeed, for a large organization I think the hybrid model most likely to succeed.
If you are centralizing some or all of your web measurement resources, then the next decision is where they should go – what’s the best chain of command? We see two very common decision paths here: placing measurement within the IT Department and placing it in the Marketing Department.
Each has advantages and drawbacks. Most organizations started by placing web sites as a whole in the IT Department. That’s almost always been changed. Now, hosting and development may reside there – but ownership is usually in Marketing (though this depends on the function of the website). Web Measurement has often followed a similar track.
Some of the drawbacks to putting web measurement in IT are pretty obvious – and similar to the drawbacks of centralizing IT. It’s notorious that IT departments lack understanding and sympathy for key business issues – so putting your web measurement here risks completely isolating your analysis from business strategy. Very dangerous.
On the other hand, there are some pretty good reasons why IT is chosen. First, as a culture, web measurement people tend to be comfortable in IT type organizations. They are technical or quasi-technical folks and they tend to work in similar paradigms. Second, almost all web measurement takes at least some IT knowledge – and where data integration issues are central, integration can be 80% of what’s involved in doing analytics. When that’s the case, it really makes pretty good sense to have your web measurement folks in the area where they can work with and get what they need from the folks who will be doing the integrations.
Putting web measurement into marketing, on the other hand, usually means that it is closer to the business function. Analysts are more likely to understand what’s needed by the business and more likely to direct analysis appropriately. So what are the drawbacks? First, most web sites aren’t all marketing oriented. There are plenty of sites where Support and Operations functions are completely slighted because Marketing owns the web measurement. That’s true even where Support and Operations are clearly more important in terms of driving web site ROI. If IT owned the measurement, it would almost certainly result in a more balanced approach. I’ve also seen plenty of situations where the web measurement department – being outside of IT – didn’t get much support for necessary integrations. And, in turn, was often completely insensitive to very real issues about data quality, data standardization and data integration.
For either direction, I believe a third issue is the fact that neither IT nor Marketing are particularly strategic. For IT, it’s probably obvious why I say this. But with Marketing Departments, it might be less so. In my view, putting web measurement under Marketing in most large companies will insure that analysts focus on key marketing concerns: particularly campaign effectiveness. That’s not bad, because these are issues that have high importance and are very amenable to web analytics. On the other hand, I think this focus tends to be too absorbing and results in web analytics getting "typecast" as campaign analysis.
So do I have a third alternative? Possibly. I think companies should consider attaching measurement to the Chief Strategy Officer. Or, to think about it differently, if I were CEO of a large company I’d attach web measurement as a direct line report to the C-Officer who I thought was most inclined and able to use analytic data. This might be the CMO, it might be the CSO, it might be the CTO. I’ll pick the CSO by default, because the CSO is uniquely positioned away from day-to-day concerns and, I hope, is directly concerned with business strategy. In addition, the CSO is unlikely to have a vast organization to manage - a problem for the both the CMO and CTO - so the web measurement will get much closer attention. I’ve stated repeatedly that web analytics needs to be responsive to business strategy to be interesting, and what better way to insure that than by giving your strategy arm the keeping of analytics?
I won’t pretend that this doesn’t raise plenty of its own issues – and it shares many of the drawbacks I’ve mentioned about for both IT and Marketing so it’s hardly a panacea. But then, there is no such thing as a perfect organization. Putting measurement at the Strategic level is very nice way to give your strategy teeth – it can arm a good CSO with the data to drive change and put tremendous pressure on organizations. Senior Executives often need a nice big club to get change. Standardized Measurement targeted at key performance issues/goals can often provide that. It’s also a great way to insure that the web measurement you do is driven by a set of questions that actually matter to the senior officers in the business. That’s vitally important if you want to get more than just another round of pretty reports from your investment in measurement.
Where you put your centralized measurement department is ultimately going to be have a lot of influence on what kind of use and value it provides. If your main concerns are standardizing reporting across many organizations and facilitating core data integrations, then an IT home may well be your best bet. If you want to insure strong business direction at the tactical level, then putting web measurement in Marketing is a good choice. If you want your web measurement to be strategic – an instrument of cultural and corporate change – then consider housing it at the CSO. No matter which direction you choose, you should be aware of the liabilities at least as much as the disadvantages – because it’s the weaknesses in your basic approach that you’ll have to work to address.
Your post is quite symbolic of what's going on in interactive marketing. I agree with you about IT, and the risk of Marketing reducing WA to campaign reporting. WA then seems to be hanging loose without a home. I don't think the CSO is the solution. First, an organization would need to have one in the first place. And I think that person would be too far away from the very concrete benefits of WA: the day-to-day analysis and optimization done on very specific elements.
I think the solution rather resides in a change of culture in Marketing. Without dimishing the role of branding, I strongly believe that interactive marketing is, or should be, the paramount example of "management by measurement", that if you can't handle quantitative analysis, or at least aren't using some of the benefits of WA, well, you shouldn't have that job.
Marketing should be the home of WA. It needs not only to give WA its rightful place, but even to let itself be transformed by it.
I think we will witness that transformation in the coming 3, 4 years. With all that's being said and written about WA and its importance, there will soon be a day when if you don't master it, you won't be entitled to the Web Manager position.
Posted by: Jacques Warren | March 05, 2007 at 07:27 AM