---

  Data Mining

By Cindy Estis Green

Carefree Resorts wanted to invite 1,000 Harley-Davidson riders to sign up for an upscale road rally. Carefree contacted a specialty list vendor to get a list of Harley motorcycle owners. They tapped their international marketing database first to match their top Zip codes with the Harley list. They then checked the demographics of those Zip codes to see which ones had a significant share of household income levels above $125,000 to match their guest profiles. Finally, they chose the highest income Zip codes with the greatest number of Harley owners. Once the lists were sorted these three ways, Carefree selected the top Zips that added up to the most qualified 1,000 Harley owners.

"The road rally was a resounding success," said Bill Gamble, vice president of sales. "Now that I have data mining technology available, I wouldn't approach a promotion like this without it."

There is a new vocabulary used by marketers: Data warehouses - large storage receptacles for data.

Database marketing - a program that takes customer information and uses it to get or keep customers. The hospitality version of this technology allows for extensive analysis of guest history, reservation databases, guest comment cards, brouchure requests and other sources of guest data.

Data mining - the process by which you distill your target customer groups out of the massive data files typically found in hospitality operation. For example, you have limited marketing resources, and you want to contact the smallest number of prospects and customers to get the greatest return. This is not an unrealistic goal, but it is not simple and will cost you both time and money.

How do you use data mining tools?

• Find out who purchases products from you so you can sell more services to those who already have exhibited an interest in your type of products.
• Create recognition and retention programs to keep your best customers.
• Improve relationships with customers.
• Identify and track promising markets.
• Help focus product development plans to match customer feedback.
• Improve how effectively your sales staff is deployed.

A basic premise of data mining is that all guests are not equal in the allocation of marketing resources. You should spend more money against those most likely to generate more revenue.

If all this sounds good, be prepared for a serious commitment. To be used effectively, database marketing needs to become a way of life for your business. It is not something done occasionally. Databases need constant attention to be sure data is current and you are not suffering the symptoms of "garbage in / garbage out" with poor procedures related to the data.

A major hotel was certain they were getting good home addresses on their guests. They put a lot of effort using post office - certified tools to standardize the address and update those guests who had moved, "household" the data (combing multiple stays under one guest name) and other typical cleaning processes.

When a guest checked into the hotel, they were asked to fill in or check their address. When asked if the desk clerk wanted a home or business address, the clerks would routinely indicate that "it doesn't matter…whichever you prefer to give." This lapse in procedure undermines the thousands of dollars spent in maintaining what management thought was a clean database.

Home addresses are the only link a hotel has to getting outside census demographics, psychographics, auto registration information and other lifestyle information to help profile its customers. It is the only means to getting accurate statistics on the most critical variable in the marketer's arsenal: geography. Knowing where your customers come from is always the first building block in a data mining program.

There is no one step answer. Once you uncover the first clue about where to find customers, it leads you to the next question until you find out enough to embark on marketing campaigns that will address your need periods.

Typically, the hospitality industry marketer starts out by getting overall profiles of customers. For the transient guests, you can find out age, income, occupation, marital status and other typical demographics. What products they purchase, arrival/departure patterns they have, what season they visit, how much they spend, and how often they travel are other commonly asked questions.

For the group guests, industry type meeting size, location, function space requirements, arrival/departure patterns and the lead time are other types of profiling questions answered with data minting tools.

After basic profiling, the marketer has to carefully guide the data - mining process. Good judgment, combined with a data mining toolbox, can yield powerful results. A database rich in good, clean information can sent you in many directions.

Management must be cold and calculating in choosing the periods that the database will address. The database cannot tell you which communication tools to use, but it can help identify new markets to test, help narrow the list of customers selected for promotional programs and help you determine whether your marketing campaigns worked.

Mapping and profiling

A small airport property outside of Washington had completed a major renovation and was reopening with a different flag. It had aspirations of average rates between $20 and $25 above its previous best.

First, a map was designed showing a 10 mile radius around the property and a list of 1,000 corporate accounts was purchased, including headquarters and branch offices with 20 employees and higher, excluding retail businesses. A round of pre-qualifying calls was made to these accounts to result in 400 qualified accounts.

This list was run against the property's previous account list of 750 accounts. There was almost no duplication in the new list. The property did sales calls on the 400 qualified accounts and closed deals on 15 in the first two weeks to generate contracts for well over 2,000 roomnights at average rates in excess of $20 higher than before. Mission accomplished - thanks to data-mining technology.

A well known resort property wanted to test market a new merchandise catalog to sell branded products. They thought the most likely customers for logo products were those who had stayed before, but their past guest list included almost 100,000 guest names and they only wanted to test a subset of 10,000 names.

Recency, frequency and value

They stratified their guest list into five segments, based on how recently the guests had stayed, how often they visited and how much they spent. Each guest was given a score from 1 to 5, based on each variable (recency, frequency, and monetary value). They excluded those who had the lowest scores for each variable and reduced the list to 250,000.

The list had been previously coded with the likelihood of responding to direct mail: this information was purchased as part of a package of demographics appended to the file. Once the direct-mail responsiveness was factored in, the final 10,000 were quickly selected. Those who actually purchase from the catalog will then be identified by which level of RFM they represent. All future lists can be further tailored to only choose those RFM levels most likely to respond.

A property did a mailing to those guests who had stayed in the last year, and offered a special summer deal. Upon analysis of those who accepted the offer, it was clear that the guests who had come most often in the past (not necessarily those who had visited in the summer) were the ones who responded to the offer. This will help the property expand what had been a practice of sending offers only to those who had visited before in the same season.

It used to be that hotels wanted to do database marketing so they could sent mailings for the holiday season to guests who came the last holiday season. It is becoming clearer to the industry that a database can be driving force for deployment of all mailing resources and the report card of results.

Data is a road map, and it is the history book of all actions. It is the guide for the product development and the crystal ball for new business development. Whoever uses their information better in the next millennium will certainly be declared the winner in the race to driving revenue.
 

For furthur information contact:

• Return to Driving Revenue  Special Reports and Articles
• Search Hotel Online