15 - Making it Work by Bringing it all Together
Thus far, the author has spoken a lot about the various ways of collecting data. Even if you do all of them, what you have is a big pile of data. The analogy the author draws (and it's quite a good one) is that you have assembled a haystack in which to go looking for needles.
Data Warehousing
The term "data warehousing" has been in fashion for some time - the concept of a company that brings together all the data in various sources and moves it to a central repository, where it can be accessed by all. In shops that do data analysis well, about 75% of 85% of the effort is expended merely to prepare the data: to collect it, bring it into a central database, and clean it up in preparation for analysis.
And once that money is spent, the data warehouse is merely a dump: it contains both useful information and useless information, and it still requires significant effort to dig up what is needed, if it can be found.
Site Critiques
Web site critiques are often sought as a shortcut: rather than analyzing performance and revenues to determine whether a site is successful, companies will hire an outside consultant to write a critique of their site, assessing it by various generic criteria.
The problem with that should be obvious: most of these criteria are based on standards that have nothing to do with the needs of the business or its customers. As a result, the critique can point you in the wrong direction, toward modifying your site to "improve" against these arbitrary standards and, at the same time, failing to meet the goals of your business.
The Multichannel Migraine
A company that is a pure-play Web retailer can limit the scope of its metrics to the Web site, but in most case, a Web site is operated by a business that is established in other channels, and the Web is often used as a medium to stimulate business through these other channels, which adds another level fo complexity.
Even when it is not your intent, the Web affects offline business. For example, return to the earlier bit about types of customers, there are individuals who will gather product and merchant information online but buy offline. How much of your in-store business is driven by the Web?
It may be significant: A survey conducted by Jupiter Media fount that 70% of shoppers would be less likely to visit a retail store if they were dissatisfied with their experience of the company's Web site.
In a case-study, IBM analyzed the retail sales of JC Penney, and found that:
EN: This has often been dismissed (it was assumed these are more affluent customers who would have bought more anyway). However, the same is seen to a lesser degree when a retailer has both a mail-order catalog and a store versus those with stores only.
Measuring multichannel influence can be difficult: there is no way to tell if a person who comes to a store for an event was driven there by advertising on the Web (any more than you can tell how many people were driven to the store by a radio spot), and because the Web is used as part of multi-channel promotions, its influence cannot be isolated.
You can attempt regression analysis or surveying to determine the influences cross-channel, but presently, there is not a good, reliable method for gauging the influence across channels.
Connect the Dots
A bit of advice: don't try to measure everything at once. For that matter, don't expect that everything can be measured. Decide what measurements matter, determine what you can feasibly do, and do the best you can.
To begin, capture basic data: visitors, sessions, click streams, and online sales. Then, move on to more sophisticated or niche-oriented metrics as you need them.
It may be that you conduct analyses as needed, different kinds at different times, and have to "connect the dots" between these individual, sporadic reports until you are able to do more regular, more extensive reporting.
You will never have all the data you could ever get out of it. But if you get what you need, and use what you get wisely, it can have a dramatic impact.
Tomorrowland
The author speaks of "tomorrow" not as a futuristic vision for the industry, but as the improvements that can result from the effective use of Web metrics.
It's a pep rally, to be sure, but much of what he has to say hearkens back to what was said earlier in the book: using the proper metrics gives you the proper knowledge, enables you to take the proper action, enables you to move your business in the proper direction.