Section 7.1. Hacks 91100: Introduction

7.1. Hacks 91100: Introduction

Now that you've read a seemingly endless number of hacks describing the data and what you should do with it, all that is left is to communicate the value of that data to your organization. Should be a piece of cake, right?

Then why do so few organizations actually share their web data throughout the company?

Data collected by JupiterResearch indicates that most companies don't share the data they collect about their successes online broadly or frequently enough. By broadly, I mean with enough people in the company; it's not enough to have a small group of webmasters look at the data; you have to get everyone thinking about how the web site is being used to drive business success. In terms of frequency, only 32 percent of companies look at their web data on a daily or weekly basis, while half that number look at the data on an ad hoc basis.

You will be successful with your investment in web site measurement only if you're able to get people throughout your company to care about the data. To do this, you have to massage the data into a format they'll understand, present it in a way that speaks directly to their interest in the Internet, and support it with training and higher-order analysis.

7.1.1. Sometimes It's How You Say It

In terms of translating the data into a format that people will understand, perhaps the best analogy is the stock market. Anyone who is interested in playing the market has access to vast volumes of information about every publicly held company that detail how they spend and make money, where the risks are, who the key players are, who they compete with and so on.

So why do people pay for market research services and visit sites like Motley Fool?

Because SEC filings are ridiculously difficult to read and because people like to have complicated information translated into more palatable terms. Unless you're in the securities industry, when is the last time you read a prospectus cover to cover? Compare that to the last time you asked a friend for a stock tip? Your friend's information may not be as complete or correct as a prospectus, but he will provide information in easily understood language, not financial mumbo-jumbo.

You must do the same thing with your web data. You have to become adept at translating the ugly details into something that anyone in your organization will be able to understand. You need to make use of relevant graphs and tablesnot necessarily the figures that your vendor provides via their application. Plan on using the language that your company uses, not trying to get everyone to learn how to "talk the talk."

The more you do to make the data palatable, the more likely the data will be consumed. The goal is to create an organization so focused on web data that people look forward to your reports, because they know you'll be providing information they can use to be more successful. While you may never experience that particular level of nirvana, if you're able to simply get people to understand and act on the data, you're winning both the battle and the war.

7.1.2. Patience, Patience, Patience

In reality, of course, none of this will happen overnight. People, for the most part, are afraid of complex data. Most people's comfort level with large spreadsheets of numbers is usually very low, unless you're a CPA or, god forbid, a professional data analyst. Keep this in mind as you build and begin to distribute your key performance indicator reports and expect that people will adopt your reports more slowly than you'd like, even if you're producing absolutely brilliant reports.

Every few months, reread the hack on [Hack #91] and use that to set and reset your expectations. Plan on supporting internal data consumers (and don't call them that to their faces, it makes them mad) as much as necessary in the first 90 days, making sure that the KPIs and reports both make sense and are being used. Go as far as cornering people in the hall and asking them about some great insight you provided in the daily KPI reportif you're really into this stuff, people already have you pegged as a geek, so what can it hurt?

The work you do early on will pay you back a thousand-fold down the road when you've managed to transform your organization into one full of data consumers, just waiting to see how the latest round of improvements have positively contributed to the bottom line.

7.1.3. Microsoft Excel: the World's Most Popular Analytics Application

An important thing to consider when you're thinking about how to distribute the data is the need to provide reports in a format that folks are comfortable with. While the application vendors have all done a pretty good job on prettifying their applications, the chances that you'll ever get everyone who needs to use this data to log into the system and successfully navigate to the right reports is, well, zero. This is why I always recommend Microsoft Excel.

Excel is hand's down, by far, the world's most popular and dominant analytics application; sort of the Shaquille O'Neal of number crunching apps. Anyone who has ever used a computer has come in contact with it, and for the most part, business people are used to using Excel for any number of things. Number crunching, project management, data sharing; you name it and I bet Excel has been used. It's even being used to analyze the behavior of Halo 2 players in the videogame world.

Given the choice between forcing people to learn a complex proprietary application or giving them the option of using something they already know and are comfortable with, most people opt for the latter. You should support them in their choice. Plan on building your key performance indicator reports in Excel and look into whether your vendor offers direct-to-Excel export or automated Excel-based reporting (a number, including WebSide-Story, WebTrends, Fireclick, and Omniture, currently do).

Excel offers an additional feature that turns out to be nearly critical to most data analysis teams: the ability to easily annotate the data, providing necessary insight. Think about this as the "chicken un-beheader"if you send everyone a spreadsheet containing really bad news, the first thing that happens is everyone runs around like chickens with their heads cut off. Bad news often creates blind panic and blind panic is bad. Now assume that when you present the really bad news that you annotate the report letting everyone know that everything is under control and you expect the problem to be corrected quickly. Voilà! The chickens don't lose their heads, and its business as usual; hence, the chicken "un-beheader."

Again, because this data is nominally complex and almost always new to folks, you need to plan on doing everything possible to increase people's comfort level. Don't reinvent the wheel; use the one you already have.