By Bruce McCracken
Though many dot-coms evaporated like raindrops from a summer shower, new media still works for some as their business continues to grow. The challenge becomes how to consolidate the growth and turn visitors into loyal customers. The questions: Who are these people and what do they want? The answer comes via business intelligence through data mining/Web analytics.
Successfully emerging from the dot-com bust, Allrecipes of Seattle, WA has become the world's number one food and meal planning site. Since the launch of their Web sites as allrecipes.com in 1997, Allrecipes offers visitors over 20,000 free recipes with sponsorship from some of the largest packaged food companies.
Esmee Williams, vice president of marketing for Allrecipes explains the business of the dot-com. "We cater to consumers in planning meals from gourmet to home cooking, all submitted by home cooks. We have millions upon millions of folks coming onto our Web site each month looking for recipes. In December 2001, our site had 4.65 million unique visitors."
With such a gargantuan flow of traffic, tracking the visitors and understanding their behaviors is akin to doing so with migrating fish in the Pacific Ocean. Business intelligence comes through what is known as data mining or Web analytics. Guy Creese, research director, Internet analytics for the Aberdeen Group of Boston, MA in the June 2002 report, "Web Analytics: Making Business Sense of Online Behavior," sums it up: "Web analytics is the monitoring and reporting of Web site usage so that enterprises can better understand the complex interactions between Web visitor and Web site offers. The resulting insight enables enterprises to optimize the site - as well as other online and offline channels - for increased customer loyalty and sales."
Essentially, the huge volume of Web data is gathered, stored, and sorted by analytic reports, as well as a series of questions or equation formulas known as data mining algorithms. Differentiation develops as does patterns that become customer affinities and segments. In this way, an enterprise is no longer boggled by not seeing the forest through the trees. The types of trees with their respective attributes and characteristics identify the forest.
Stumbles and Grumbles
With limited staff, infrastructure and budget, Allrecipes took an awkward and futile stab at data mining. As a cost savings, Allrecipes bought a software package and had only one person dedicated to the task at hand. According to Williams, "What made it so difficult is that we are actually 27 different sites that appear to be one to the customer. So that when we went to run the logs, we have 27 different sets of data to run which make it an impossible task. It would take a week. By the time it was done, it would have to be started again. So we were able to run the log files but never extract any data from them. It was sort of a ridiculous task."
Subsequently data was gathered with no resulting information of value says Williams. "There wasn't much in there that marketing could make much out of. We knew that sales and marketing needed some data that we weren't able to process. We decided to outsource, as running the logs internally was becoming unmanageable to get the results that we wanted in the time that we needed." The process had resulted in stale leftovers for IT and indigestion for sales and marketing.
In April 2000, Allrecipes chose digiMine of Bellevue, WA, founded in 2000. The service provider gathers data from the client's system and integrates the information into a hosted data warehouse. The data is then cleansed, parsed and filtered based upon the needs of the client to enable reports for a wide array of business intelligence about the behavior of site visitors.
Creese points out the advantages for a firm like Allrecipes outsourcing to digiMine. "DigiMine offerings will be attractive to companies that want access to sophisticated and customizable Web analytics, but do not want to pay the personnel, hardware, and software costs of an equivalent in-house solution. DigiMine focuses on helping enterprises quickly answer unique questions - a form of personalized service that very few of its competitors offer."
The comparative differences in buying software versus using an application service provider (ASP) like digiMine in the table below from the Aberdeen report show the advantage of an ASP for an enterprise like Allrecipes.
Software Licensing Model Versus ASP Model
|Location||Software within the enterprise||Hosted at the ASP's site|
|Installation Speed||Weeks or months||Days or weeks|
|Pricing model||Software license fee up-front||Monthly fee|
|Price range||$500 to 1 million +||$3,000 - $300,000 per year|
|Accounting impact||Server is a depreciated asset on the balance sheet||Server fee is book as an operating expense|
|Incorporates data from offline channels||Often||Sometimes|
|Site monitoring||Monitors one to several sites||Monitors one to many sites|
|Customization capabilities||High||Low to high|
|Source: Aberdeen Group, June 2002|
A major problem for many smaller companies in implementing an in-house business intelligence solution like data mining is that the task falls upon IT even though the users will often be sales and marketing. Frequently, the IT staff does not understand the business needs. Conversely, sales and marketing often does not understand the technology.
This gap is bridged, according to Joel Sider, manager of corporate communications for digiMine. "Even if you get a system in place, if it doesn't speak the language of the business user and make sense, they can't apply it. Our account management team works with the client to ensure that we meet the needs of the business user while working within the constraints and policies of their IT department."
Deployment was completed by June 2000 and was a painless process for Williams. "They were able to make the solution so easy for us. It didn't require much to take all of the tasks away from us. They are very bright, very driven and entrepreneurial."
Seeing With Vision
Allrecipes immediately began to see their business with increased clarity according to Williams. "We discovered that we had less retention, that up to 50 percent of our traffic is new visitors. We had more churn then we thought. Our audience is largely female and research suggests that females are much more solution based on the Internet. They want to come in, find what they want to meet their immediate needs." Through understanding these visitors, Allrecipes has been able meet these needs, leading to return visits.
The knowledge enables Williams to be proactive. "We can now entice more folks to come back more frequently. DigiMine enables us to look at the top key words in consumer searches so we can tell what ingredients or cooking styles are of interest at a particular time. We are able to craft our editorials to hit on subjects that were key and of immediate interest to the consumers. We know exactly what our users are looking for, what features they are gravitating towards. The visitor pattern has become more frequent so they become much more loyal users."
Williams says that Allrecipes is a happier home for the staff. "Outsourcing to digiMine has made our IT staff happy as it keeps sales and marketing out of their hair. Sales and marketing can now access digiMine where ever and whenever they want to gather information and it doesn't put anyone out. Everybody is able to focus on his or her jobs more with less interruption."
The ROI and growth is synergistic and stunning, as Williams reports. "We would need at least three to four people in-house to mirror what they do and we wouldn't get near the results. We are probably saving anywhere from 300 to 400 percent."
Add to that, a successful growing business with the help of the expertise of digiMine that leaves Williams satisfied. "DigiMine has been able to keep up with our growth and add refinements. We have tripled our traffic in the last two years." Visitors are becoming customers as it is just like home cooking. If they like what they taste, they will come back for second helpings.
Bruce McCracken is a business writer with specialization in outsourcing. His coverage areas are primarily in IT, eCommerce, CRM, HR, and supply chain/distribution with focus on small to mid-sized companies. His work, useful links, and commentaries with guests may be seen at www.brucemccracken.com. He may be emailed at firstname.lastname@example.org.