jim.shamlin.com

The Web Analyst Tackles Multichannel Metrics Online

The Internet is itself a multichannel environment. While many companies focuses solely on the performance of their Web site (with special emphasis on the online storefront), customer communication also takes place on associated sites, e-mail, online advertising, online publications, social networks, blogs, and chat rooms.

Website Analytics

Web site analytics are "surprisingly unappreciated" at most companies. They are seldom available to others within the company, and are disregarded even when they are published throughout a company. They are often seen as pertaining only to the operation of the online storefront, but they have implications for marketing in other media, as well as the operations of the entire organization.

Web site analytics indicate how customers are attracted to the site, which ones buy, which products they buy, how often they buy, what marketing messages are most effective in reaching the best customers, what other information the individual considers. If you divorce this information from the channel, it represents the kind of information that would be considered "gold" to marketers in any other medium.

Some of the insights that Web analytics provides are:

Much more information is available for close analysis of Web site usage, and this information is valuable in other media as well.

Web Analytics Primer for Multichannel Marketers

EN: Since I've site analytics and developed custom solutions, I'll elide the details of much of this section - it'll be choppy:

Off-the-shelf solutions are not the best approach. They analyze separate phenomena (one bit of software checks click-trails, another analyzes search queries) in a disintegrated manner, and seldom coordinate this data to meaningful behavior.

When it comes to Web pages, the software collects data on hits. Hits can be clumped into visits by checking the referring site IP. Visits can be clumped into visitors by putting a cookie on the remote user's system. Visitors can be turned into people by having them register a unique username and log into the site.

Traditionally, the "big three" metrics have been unique visitors (number of people who come to a site), sessions (number of times a visitor visits a site), and page views (number of pages a visitor sees over time). There are some attempts to create industry-standard definitions and methods for analyzing the data.

However, not everything that can be measured is of importance: you must decide what is relevant to your business and define key performance indicators (KPIs) accordingly.

It's also noted that most commercial solutions do not handle segmentation well (or at all), and it is especially important to sift through the amalgamated numbers to track specific segments (new customers versus established ones, the "top" customers in terms of visit frequency, etc.)

And finally, keep in mind that your analytics should be actionable. If you don't intend to do anything with the data, collecting and analyzing it is a waste of time and resources.

Attribute Responses to the Right Online Marketing Efforts

It is difficult to assert causation for customer behavior. If there is an increase in sales, was it because of a print advertising campaign, a favorable magazine review, display advertising at a sporting event, the results of increased word-of-mouth from a popular blog, or just happenstance?

Online, it may be possible to determine, with some degree of accuracy: the site/message that referred users to a specific page may be tracked by a referring address, a specific landing page on your own site, or query string data attached to the link.

Tracking codes, which identify a specific offer, can also be used online (similar to codes embedded in retail coupons). On the Web, this is generally done through query string data, and may include code that identifies the particular customer as well as the source that sent them (when marketing to known individuals).

Ad-hoc analytics can also be run for a specific campaign, identifying and perhaps isolating behavior that resulted from a specific action, either immediately (a person clicked an ad from a promotional e-mail and bought a product) or long-term (a person who received three specific messages over the past month visited the site).

A few notes of caution about outsourcing: some third-party services that help with promotion track the performance of their own campaigns in a way that can interfere with your ability to do your own tracking. Also, be aware that the information they collect about your customers may be made available to your competitors.

The use of micro sites (separate Web sites other that the main one) for ad campaigns is mentioned. On the upside, this gives you the ability to present a highly specialized message to the intended audience without affecting what other users see, and to isolate behavior resulting from the campaign. On the down side, the individuals reached by such a campaign do not identify the company's actual Web site, and may be lost when the campaign ends and the micro site is torn down.

Success Metrics for Online Marketing

Two separate sets of metrics are suggested: one for the short-term performance of a marketing campaign, and another for the long-term goal of customer retention.

Campaign-Centric Metrics

There are a number of different approaches to measuring the success of a campaign, but the most important is ROI, which measures the revenue received as compared to the cost of conducting the campaign.

Other measures can be used to analyze performance to determine where drop-off occurred between the time the user was shown an ad and the time they bought (or the place on the path to purchase where they dropped off)

Customer-Centric Metrics

Whereas the campaign metrics look at sales created by a campaign, the customer-centric metrics look at how an individual interacts over time.

A "promotion history" shows which campaign a given customer was exposed to over time, since their initial acquisition


Contents