What does the future hold for web analytics, indeed? During my tenure at JupiterResearch I was more-or-less paid to predict the future, but at best I was right maybe 50% of the time. I think I predicted (correctly) that Google would give Urchin away, but I probably also predicted that Microsoft would acquire WebTrends (incorrectly) as a result. Such is life.

Fortunately I am smart enough to surround myself with really smart people, which is what I have done at the newest Web Analytics Demystified weblog: The Future of Web Analytics, Demystified.

What Joseph Carrabis (my partner) and I are doing at “The Future of” blog is creating an opportunity for some of the brightest voices in our community to wax philosophical about where we’re all going and what things will look like when we get there. In the last two months we’ve had excellent conversations started by the likes of Joseph Carrabis (NextStageEvolution), Rene Dechamps Otamendi (OX2), and most recently Mr. Ian Thomas of Microsoft fame.

The unique thing about this web analytics blog is that posts and comments are basically peer; we’re looking for long, well thought out comments that add something to the conversation. And as odd as it sounds we’re not approving navel gazing, fawning, and trackback/ping so that we can keep the conversation moving!

If you’re a long-time reader of my personal weblog I would strongly encourage you to subscribe to The Future of Web Analytics, Demystified. More importantly, if you’ve got a big brain and want to help us work collectively in an effort to figure the future out before we get there, we welcome your comments. Alternatively, if you have a prediction, see a problem, or want help resolving a problem that you struggle with, I’d love to hear from you about being an author in the Future Collective.

(Last week I published PART I of this post which you should read first if you haven’t already done so.)

STEP FOUR: DETERMINE YOUR KEY PERFORMANCE INDICATORS AND CRITICAL REPORTS

You’re probably thinking “shouldn’t we have done this after we defined our business objectives and activities?” Conventional wisdom would probably say you should, but in my experience if you don’t have a clear process for leveraging those key performance indicators (KPIs) and critical reports, you may end up with one of three things:

1. A huge report of 40 KPIs distributed across the organization that few people are likely to read and even fewer likely to act upon
2. No KPIs distributed at all, and the expectation that everyone will simply “log in” and get the information on their own
3. Well-defined and clearly articulated KPIs distributed hierarchically throughout the organization (because hey maybe you read a great book on the subject at some point)

The problem is that only the third possibility will deeply benefit your organization. I know that some people talk about hundreds of internal users who really get web analytics and all make superb decisions with the data, but this is very much the exception, not the rule. Remember, in our Web Analytics Demystified Spring Survey 69 percent of respondents said that they did not believe the majority of people using web analytics data in their organization actually understood that data.

It is far better for your analytics hub, as mandated by their executive sponsor in agreement with his or her peers throughout the organization, work directly with the individual spokes to ensure that appropriate KPIs are defined and the basis for those measures is clear. The hub then follows-up with appropriate explanation about the measures, including training on the reports and data that forms the basis of the indicators.

Your critical reports are directly tied to your key performance indicators (which remember are tied directly to your business objectives.) If you belong to the marketing organization than your KPIs will be measures like “Campaign Response Rate”, “Campaign Conversion Rate” and “Campaign Cost per Click”. Obviously as these KPIs change, appropriate tactical resources in the marketing spoke will review campaign response, conversion, and cost reports in your analytics application.

Your KPIs and critical reports will differ dramatically depending on what department you work for and where in that department you work — remember that the best practice for key performance indicator distribution is to deliver the specifically and hierarchically. Most attempts that I have seen to send “everything to everybody” have failed (often miserably).

STEP 5: DETERMINE HOW YOU’LL DELIVER ANALYSIS

Once you know what your KPIs and critical reports will look like, the next step is to determine how you’ll produce and deliver analysis. Let’s assume for a moment that you’ve got a hub-and-spoke model in place and the hub is receiving regular requests for more information, insights, and recommendations. The question then becomes “how will you deliver those insights and recommendations?”

As I said last week, there is no one “right” way to communicate about web analytics data but there are many, many wrong ways. The central challenge when delivering analysis stems from the fact that so few people really understand what web analytics terms mean, what the limitations of the technology are, and what is possible and impossible to report on. But it’s not like you can just give up and ignore the confusion, so what’s a great analyst to do?

The answer is “work harder, and think outside the box” (to use an overused term). While reports and raw data are best delivered using the Bottom Line Up Front (BLUF) method, analysis really needs to be more engaging. Remember: when you deliver analysis, what you really need to do is to convince the listeners that they need to take some action. To do this you absolutely have to be engaging.

Things that have worked for clients of mine in the past include:

• Well-delivered presentations, given IN PERSON, not just sent via email in hopes that people will review and understand
• Well-written documents, followed by a meeting to make sure that everyone READ the document and is on the same page
• Short summary documents, written up like a newsletter or newspaper article, designed to get people to attend a meeting or presentation

Since we’re in a Web 2.0 world, and since many of you are increasingly comfortable using new technology, a few other things you may want to consider include:

• An internal analysis Wiki that people can subscribe to and participate in. The Wiki is a good idea because it allows you to capture the conversation in a searchable format
• A regular analysis podcast, providing an update on past analysis and summarizing the data currently being reviewed
• A analysis video or vidcast, created with tools like TechSmith Camasis that allow you to easily blend images, live screen capture (useful when showing people live data in your analytics application), and annotation

The advantage the final two ideas confer is their ability to be downloaded to an MP3 player like the iPod or iPhone. If you have busy executives, you might be better able to reach them if you give them something to watch on the airplane or listen to on the drive home.

Keep in mind that none of these “Web 2.0″ strategies should replace well-written, well-presented analysis, delivered in person whenever possible and making specific recommendations for changes (including a testing plan when possible!)

STEP 6: PUT IT ALL TOGETHER!

Assuming you’ve completed the previous five steps, you now have a functional web analytics organization, one capable of delivering relevant reports and producing actionable analysis. Now the challenge is to stop spending all of your time generating reports and start delivering analysis!

Unfortunately, for many organizations this is really, really difficult. Even when there are dedicated resources — people specifically hired to do web “analytics” (not web “reporting”) — far too many bright folks end us spending all of their time churning out reports. Even worse, these reports often go unread, unused, and unnoticed despite the real and opportunity costs associated with generating them.

To be really, really successful with web analytics you have to train the organization to stop looking for reports and start asking for analysis, insights, and recommendations. While every situation is different, ask yourself how closely your organization follows these steps:

1. Automated KPI reports arrive, highlighting a potential problem associated with a core business objective
2. Line of business analytics resources consult critical reports directly looking for a reasonable explanation
3. Failing a reasonable explanation, business resources request analysis resources from the analytics hub
4. Analytics hub double-checks LOB’s cursory analysis, confirming the need for deeper exploration
5. Analytics hub prioritizes analysis with the business based on pre-agreed criteria
6. Analysis is delivered back to the business along with recommendations and a testing plan
7. Recommendations are reviewed by the business, test plan is agreed upon
8. Tests are run, results are socialized as follow-up to the original analysis
9. Incremental value of change is recorded to help calculate web analytics return on investment

Individual departments are still getting their reports, but they’re generating them by themselves. Senior managers have an appropriate view into the metrics, and their own resources to evaluate observed changes. Those resources have a way to get help when help is needed. Help (the hub) isn’t bogged down generating ad hoc reports all the time and is able to focus on high-value priorities. People produce analysis and make recommendations. Recommendations are tested. Optimization happens.

Kinda brings a tear to your eye, doesn’t it?

I know there are a hundred other things that come up in the line of business for any of you who are working practitioners, but having a clear communication strategy is the first step towards whittling that list down to something reasonable and, more importantly, valuable to your organization. Defining your business objectives, clarifying ownership and organization structures, establishing KPIs and critical reports, and knowing what your analysis output will actually look like is fundamental.

Defining your web analytics communication strategy will let the data work for you, not make you work for the data. It will help you move from making purely tactical decisions and start using web analytics strategically as part of your entire business. Over time you’ll find that a clear strategy, no surprise, helps the entire organization better understand web analytics in general and the value your investment can provide. And perhaps most importantly, a clear strategy will cut down on the volume of under-used, unused, and ignored reports traveling across your network.

If you’re interested in defining a web analytics communication strategy in your organization, I’d love to talk to you. If you don’t need help, I’m still happy to provide encouragement. If I can help you, great. If I can’t help you, I bet I know somebody who can!

Judah’s recent post titled “what does your web analytics team look like” reminded me of something that has been on my mind a lot since I presented my Web Analytics: A Day a Month webcast for the American Marketing Association last month. As I travel the world talking about web analytics to companies of all shapes and sizes, one thing I’m struck by is the number of differences in how companies approach sharing web analytic data and information.

It’s not as if there is any one “right” way to communicate about web analytics, but it is clear that there are many, many wrong ways to do it. But rather than dwell on wrongness, I prefer to focus on rightness so here are a few thoughts on developing a clear strategy for communicating web analytics.

This post may seem pretty basic to many of you, but if it does I would encourage you to ask yourself these questions:

• What decisions are web analytics driving in your organization?
• Are those decisions largely tactical or are they truly strategic?
• Do you feel like most people in your organization understand web data?
• Are you producing reports that are going under-used, unused, or are flat out being ignored?

If you are less than impressed with your responses I would encourage you to read on. I’m not saying you’ll necessarily learn anything new, but maybe you’ll read something that you think your boss should hear.

STEP ONE: DEFINE YOUR BUSINESS OBJECTIVES

I know, I know, you’ve heard me say this before. I’ve been saying this since 2002 but I’m going to keep on saying it since it bears repeating. By clearly defining your business objectives you get two things done:

1. You remind everyone in your organization why you have a web site and why those of you who work online come to work every day.
2. You build a framework against which you will define the core activities and interactions that are worth measuring and communicating

The second point is important: You cannot measure everything effectively and efficiently — you have to have some basis for deciding what to measure and what to report. I have seen any number of companies work hard to collect “all possible data” only to realize that few people are actually asking for that data and even fewer are doing anything with it.

When you define your business objectives and get consensus on what is most important to your online business, the measurable activities that you will be communicating across the organization become clear. Suddenly rather than struggling to measure every aspect of every page across every segment you’re able to focus on critical measures in critical paths in your most important visitor segments.

I covered all of this in Web Analytics Demystified what seems like years ago and again in Web Site Measurement Hacks (which you can now purchase direct from my site, had I mentioned that?) but again it is worth repeating. And while it is far less common now that I will ask companies about their business objectives and get conflicting opinions, many companies have still not gone through the process of clearly documenting these objectives and the associated activities to serve as basis for their measurement efforts.

STEP TWO: DETERMINE WHO OWNS ANALYTICS AT YOUR COMPANY

One of the biggest problems I see in web analytics today is a lack of clarity regarding ownership of analytics inside the organization. On this point I will be as clear as possible:

The owner of web analytics in your company NEEDS to be someone senior enough to ensure that analysis is being produced and used!

I spend an awful lot of time as a consultant talking about ownership and structure in analytics. Your executive sponsor needs to be closely connected to web analytics and have a clear understanding of the value and opportunity measurement provides. If this is not the case, you may spend an awful lot of time producing reports that go unread and analysis that goes unused.

I suspect that my fellow blogger Daniel Shields can attest to the goodness in this recommendation, working for a great boss at CableOrganizer, but more often than not when I ask the question “Who owns web analytics?” I get responses that talk about budget centers, middle-management who haven’t got budget authority or enough political clout, or worse yet, nothing but uncomfortable laughter.

Clients almost always ask “Where should web analytics live? Should it live in Finance, I.T., Marketing, or Research?” to which I almost always answer “Who is the most senior, well-connected person in your organization that is likely to really understand what web analytics is good for?” and then give their department as my answer. Here are some additional thoughts:

• Finance: Analytics living in your finance organization is fine because your CFO understands how to produce detailed analysis and make that analysis valuable internally
• Marketing: Marketing is great since in many cases marketing has the most to gain (or lose) based on web analytics data and analysis
• Research: If you have a market research organization this is also a great home since the analysis team in research usually has an excellent understanding of the customer and their (offline) behavior
• Information Technology: I personally don’t usually recommend that web analytics live in I.T. There is often too much baggage and a disconnect between I.T. and the business for this to work (but I do know of a handful of examples where I.T. ownership of analytics does work)

At the end of the day the most successful analytics organizations are those where the executive sponsor “gets it” and is able to champion for the cause at a very high level. They will need money, resources, and time from the rest of the company to deeply integrate the necessary web analytics business processes, so seniority is an absolute must.

STEP THREE: DETERMINE YOUR ANALYTICS ORGANIZATIONAL STRUCTURE

This is the step I’ve been thinking a lot about lately, how analytics organizations are structured and integrated into medium-to-large-to-very large companies. As I’m sure you know, this piece is far from a no-brainer — whether you subscribe to 10/90, 10/20/70, or some other percentage-wise distribution of effort, I think we can all agree that people are critical to web analytics success.

But as Judah deftly points out, just hiring someone is only the beginning of the work: The more important piece is determining how those resources are going to actually provide benefit back to the entire organization. You need to have a clear strategy for leveraging these resources to produce the maximum number of insights possible.

For about four yeas I have been talking about the “hub and spoke” model for web analytics organizations, especially to medium, large, and very large companies. The hub and spoke is basically a centralized/decentralized model for measurement, one that centralizes deep analysis expertise for use across the organization but mandates that each individual department and line of business takes responsibility for their own reporting needs.

The folks in the analytics hub are directly responsible for things like:

• Producing analysis, real analysis, to support business decisions
• Providing training out to the rest of the organization on tools and data
• Communicating about the goodness (or lack thereof) in the data collected
• Interfacing with the vendor(s) providing measurement software and services
• Managing multivariate tests and analyzing their results
• Working with I.T. to make changes to data collection and integration

Perhaps most importantly, the hub work directly for the executive sponsor for analytics (see STEP TWO above.) Establishing a real web analytics hub is the first thing you need to do if you want to STOP spending 80 percent of your time generating reports (something a prospect recently referred to as being a “report monkey” which they didn’t seem super-excited about …)

The folks in the individual departments and LOBs are responsible for things like:

• Paying careful attention to their key performance indicators and react to observed changes
• Spend enough time learning the available technology to answer at least basic questions when changes are observed
• Generating whatever reports are necessary on a regular basis and modifying those as required
• Interface with the analytics hub to ensure that requests for testing and analysis are clearly communicated
• Respond to test results and analysis by putting the insights generated to work for the organization

The best possible news is that the folks in the spokes don’t have to be web analytics experts! Hell, they don’t even need to read the available literature if they don’t want to (but they should.) They really only need to take enough time to learn what their KPIs are telling them and which reports in the analytics application(s) are relevant when things change.

Thinking about the relationship between the hub and spokes:

• The hub does analysis, and the spokes do reporting
• The hub executes multivariate tests, but the spokes recommend them
• The hub work directly with I.T., the spokes get to continue avoiding I.T.
• The hub helps to plan, manage, and monitor KPIs, the spokes live and die by them
• The hub runs something like Omniture Discover or IndexTools Rubix, the spokes use SiteCatalyst or Google Analytics

This is great news because there are many, many people out there that have a 0.2, a 0.33, or a 0.5 FTE for web analytics — not nearly enough time to really get deep into web analytics but enough to create the expectation that they’ll use the data to make business decisions. The hub and spoke model creates a business process to support partial FTE in their endeavor to use and benefit from web analytics, which those partial FTE seem to truly, truly appreciate!

In my experience, over time the people who really like this kind of work will pop up and ask great questions, looking to push the boundaries of their understanding of “our little craft.” They’ll read books, blogs, go to conferences, etc. and over time may realize that they really want to work in the field of web analytics full time. Which is great, because without those people flowing into the system, the multitude of recruiters and companies across the globe looking for experience web analytics professionals haven’t got a prayer.

Since Judah, Daniel, and I have been talking about the length of out posts lately I think I’ll stop here and publish Part II of this post later this week.

The key takeaways from the thoughts here are:

1. You have to have a web analytics communication strategy
2. You have to clearly define your business objectives and supporting activities
3. You need to define and establish an analytics organization
4. Your analytics organization needs to report to an appropriately senior person
5. The hub and spoke model for web analytics has many advantages, especially in large organizations
6. Web analytics done well has a tendency to make people more, not less, interested in web analytics (which is good!)

A lot has been written recently on the hot job market for web analytics positions. Perhaps because there is such a profound gap between the number of open positions and the number of experienced practitioners looking for jobs, web analytics salaries run pretty high. We collected a bunch of data on salaries earlier this year and our report on the subject is available now.

Some of the key insights:

• The average U.S. web analytics practitioner, across all job types (end-users, vendors, consultants, agencies) is reported to be earning $86,883 annually in salary and bonus
• Those of us working for vendors are the best paid, with the average vendor practitioner earning $105,192 in salary and bonus
• Unless you’re working for a very large company (5000+ employees), the difference in average salary is negligable
• There appear to be three “tenure groups” for salary, which may be as much a function that people get raises by switching jobs as they do waiting patiently for a substantial salary increase
• As we’ve published previously, companies relying on defined business processes are paying far better than those relying on employees or those lacking any formal approach towards web analytics
• There may be hope on the horizon for those companies actively seeking experienced practitioners

Because I’m trying to write shorter reports, I’ve also created a companion web page that will allow readers to examine all of the salary data we used to produce this report, segmented over 30 different ways. Hopefully everyone will find this data interesting, regardless of whether you’re looking to hire or looking for a raise. The report and companion data are available here:

• Web Analytics Salaries 2007: Insights and Observations (PDF format)
• Web Analytics Salary Data

I welcome your feedback on this research — your reaction to the data as well as your opinion about the value of our publishing this research on an annual basis. I have to admit I rather enjoy being able to publish this research so hopefully everyone out there is finding it valuable.

UPDATE: I am lame for forgetting to thank June Dershewitz, James Gardner, and Dylan Lewis for spending the time to review this report prior to its publication.  Thanks June, James, and Dylan!

Omniture just bought Visual Sciences for $394M. Congratulations to all involved. It will be very interesting to see how this plays out for clients, prospects, and the entire industry but exciting times indeed.

I sure am glad I predicted industry consolidation by the end of the decade back when I worked for JupiterResearch.