Building Marketing Databases In-House or Outside Bureau

Building Marketing Databases

In-House or Outside Bureau? • Outside Bureau: • Outside agency that specializes in designing and developing customized databases. • Three factors to consider • Cost • Control • Customization

Cost • Cost of developing a database in-house depends on – Existing workload of the in-house staff – If additional help is needed from contract programmers – Other system development initiatives that will be forgone – Whether an outside resource is quicker and better

Control • Customer list is the lifeblood of most customer centric companies. • External facilities create additional risk of exposure. • Companies need to manage risk through procedural and contractual measures. • External bureaus should have familiarity with • The company policies • Initiatives • Processing systems, and • Data

Customization • In house systems are more customized to the marketing departments needs. • However, end-user departments may be charged by the IT department on a cost plus basis. • Sometimes the cost plus charge is costlier than an outside bureau. • Service bureaus have very flexible approaches to working with several clients.

Customization • However, internal IT staff are always available to make last minute changes. • Internal IT staff may be trained to work only on standardized systems. • However, a dynamic system such as marketing requires special skills to design.

External Vendors • What if the platforms/software of marketing department and IT department do not match? • Time is spent in making the marketing and IT staff speak the same language. • Until then the marketing department uses the same archaic information to make decisions. • External vendors alleviate these problems because they have previous experience helping previous clients.

Combined Approach • Approach #1: – Use external bureaus and internal staff for tasks they are specialized in respectively. – For example, Internal staff may be good at design while the external supplier may be skilled in implementation. • Approach #2: – Use an external supplier initially, and then migrate the system to internal resources later.

What Data is Needed? • Three step process: – Identify business and promotional files that should be included. – Review data elements in each contributing file. – Select from each file the data elements required in the marketing database.

What Data is Needed? Identify files to be included: • Not practical to have all the information that ever existed about a customer. • Issue to consider: – Is the file too hard to get? – Will the database contain data on customers or prospects? – How much promotion history is available? – What levels of transaction data is available? – Is the desired data actionable? – Data element coverage.

Review of Data Elements Review Data Elements: • Detailed reviews of each data element with the data processing or end-user personnel. • May take anywhere from one day to a few weeks. Select Data Elements: • Include all the data elements identified in the review process– even somewhat critical elements. • Review the selections once a year after experience is gained with the system.

How will data get into a marketing database? • Data extraction by data providers – Standardized programs written to extract data. – Automated when updates are required. – Is not flexible when new data elements are added. • Data extraction including creation of summary calculated data • Un extracted copies of data to data loader – Provide whole data files to the loader. – Loader decides on what records to include or exclude. – More work for data loaders. – Simpler, and cost-effective when data elements change during review

In house Vs Service Bureau Implications • Data availability on a regular basis depends on – In-house • Workload of IT department (less number of data sources) • Co ordination between departments (large number of data sources) – Service Bureau: • Billing structure of the service bureau • Cost of not having to co ordinate with other departments may be steep sometimes.

Consolidating Records • Customer data is captured in transaction oriented systems resulting in duplicate records. • Data has to be cleaned before it is loaded into a database. • Example: Financial service firms – using Unique customer ID or unique account number.

Data Cleaning • Duplicate Identification – Software used to identify duplicates in mailings – Based on names and addresses • Address Standardization – Representative software: Finalist, Code-1 Plus – Verifies if 123 oak street, or 132 oak street, or 123 oat street is correct. – Uses national zip code databases for this purpose.

Data Cleaning • Logic of address standardization software - The ZIP code matches the city and state portions of the address - The corrected ZIP code contains an Oak Street or both - The street numbers along the street in question lies within bounds of the ZIP code - The street number is a valid address - Whether the address is a business or a residence

Data Cleaning • Matching Issues - Variations of spelling of an individual’s name - Companies with multiple business files store name and address in different formats - Use matching software in first case and parsing in second case - Duplicate identification may be very important for prospects too

Data Cleaning • Scrubbing - For financial services companies, registration info my be imbedded in customer record - Use specialized application of parsing and dictionary use - Software would recognize that some terms are not part of names or addresses - Scrubbing software also retains meaning of registration data

Data Cleaning • Householding - How certain are you (with the data available) that Anthony Smith and Julia Smith are married? - Householding based on names and addresses -Constitute a confidence indicator. - Mail material that promote a happy married life only if you are 100% confident.

Data Cleaning • More complex households – Type I: Same last name, same address, and joint account. – Type II: Different last name, same address, and joint account. – Type III: Common account, different addresses, may or may not have same last name. – Type IV: Same address, different last names, and different accounts.

Frequency of Update or Replacement • Frequency of updating determines how well the database represents the actual customer master file. • Updates need not be more frequent than the decision making interval. • Intervening month’s files need to be accessible to make the update processing.

Avoid Access Conflict • There is a always a priority conflict between – Analytic users of data: Marketers – Operational users of data: Customer Service • Solution is to maintain separate databases with each. • But it is too costly, marketers may not have up-todate information on some transactions. • To calculate trends, distribution by list source etc. however, real time data may not be necessary.

Other Issues • Level of Detail • Transactional Vs Summary • Update Vs Replacement • Determines analysis that can be performed • Determines variable cost • Conduct periodic reviews to determine changes in business needs and the resultant changes in the databases.