{{ searchResult.published_at | date:'d MMMM yyyy' }}

Loading ...
Loading ...

Enter a search term such as “mobile analytics” or browse our content using the filters above.

No_results

That’s not only a poor Scrabble score but we also couldn’t find any results matching “”.
Check your spelling or try broadening your search.

Logo_distressed

Sorry about this, there is a problem with our search at the moment.
Please try again later.

When Seth Romanow was the Director, eMarketing and User Experience eBusiness Systems, Global Business Solutions at Compaq Computer Corporation, he wanted to sell more computers from the website. His team systematically tackled the problem with a ten-step program.

1. Instrument Web Pages

Each web page is a stepping stone along the path to the virtual cash register. It's not critical to keep an eye on every step, but there are milestones that are worth noting. Compaq identified those milestone pages and gave each a small piece of code to record when a visitor clicked - something a bit more reliable and informative than pageviews from a server log file.

2. Collect Clickstream Data

To answer how people traversed the Compaq site required them to capture clickstream data - the sequential movement of each individual from milestone to milestone.

3. Link in Order Data

Compaq wanted to know where people went, but they particularly wanted to know how people who made a purchase went about the task. So they integrated data from their ecommerce system where the orders are placed. with the clickstream data from the website.

4. Link in Campaign Data

Which promotional campaigns (direct mail, banner ads, ads within the Compaq site itself) resulted in sales? Once Compaq recorded the effectiveness of a promotion to identify those with the greatest response (clickthrough), it was time to match them up with the clickstreams through to the sales to determine which ads resulted in the most income.

5. Link in Survey Data

Watching what people do is fascinating, but Compaq realized that, like watching wild animals, there is a limit to what you can understand about observed behavior. There comes a time when you must ask your subjects for their thoughts. Where did they have problems? What were they expecting? A series of event-driven pop-up surveys were implemented to capture that first-hand data from the customer perspective.

6. Trust MOM

With the data collected, it was time to crunch the numbers so Compaq created the Marketing Optimization Machine (MOM). MOM looked at ads within campaigns by day and calculated the effectiveness by comparing total visits, new visits, repeat visits, total purchases, number of pages viewed, average session time, average page download time, conversion rate and, well, you get the idea.

7. Tune Individual Pages

If visitors who resemble buyers get stuck on a given page and do not buy, what can be done about that page? Using a combination of dispersion data (where did they go from an individual page?) and survey data (Why did you leave?) Compaq found that some pages offered too little information, some too many choices and some were simply difficult for people who were not born and bred at Compaq to understand.

8. Optimize Navigation

The logical extension of page tuning is path tuning. Not all visitors followed the logical path from home page to product description to purchase that Compaq so carefully laid out. Some did. Some went right for the buy and some meandered about the site for more than 50 clicks. Some knew what they were after and found it readily and some bounced back and forth between the same pages. Helping these ping-pongers to become successful navigators resulted in more sales. Figuring out where people leave the buying process and plugging those holes resulted in more sales.

9. Enhance Search

Many people come to a website and enter a term in the search box at the top of the screen even if there's a big button, right in front of them, linking to what they want. Compaq recorded the most frequent searches and then studied the results the search engine produced. By manipulating those results, they dramatically increased the number of "successful" searches and, therefore, sales.

10. Optimize Merchandizing

What do people look at before they place items in the shopping cart? How can those pages be tuned to encourage more people to do so? What do people put in their carts? Which items are most frequently purchased together? With this information, the Compaq web server can be programmed to offer the spare battery to the laptop buyer, the wireless mouse and keyboard to the desktop buyer, and the extended warranty to both.

The secret to their success was assigning a value to each project, indicating the potential percent increase in revenue if each identified stumbling block was corrected. That gave them a very clear list of priorities.

How successful was their page tuning and visitor tracking effort? Let's just say that after Hewlett Packard bought Compaq, they made Seth Romanow the Director of Customer Knowledge Management and Analytics, eBusiness and Customer Operations. He's now in charge of all customer data from all touch points. I'd say that counts as successful.

-----------------------------------------------------

About Jim

Jim Sterne is an internationally known speaker on electronic marketing and customer interaction. A consultant to Fortune 500 companies and Internet entrepreneurs, Sterne focuses his twenty years in sales and marketing on measuring the value of a Web site as a medium for creating and strengthening customer relationships.

Sterne has written five books on Internet advertising, marketing and customer service including, his most recent, "Web Metrics; Proven Methods for Measuring Web Site Success." Sterne is the producer of the annual Emetrics Summit www.emetrics.org and is a Founding Director of the Web Analytics Association (http://www.webanalyticsassociation.org) www.webanalyticsassociation.org.

-----------------------------------------------------
Emetrics Summit, Web Analytics Conference
Santa Barbara, June 1 - 3
London, June 8-10
www.emetrics.org
-----------------------------------------------------

Published on: 12:00AM on 4th May 2005