A guide to data quality

Colin Rickard, managing director EMEA, DataFlux, looks at managing that most valuable of company assets...

Data quality, dangers and depreciation

What would you say your company's most valuable asset is? Is it your customer base, your property or product suite? What other assets do you have that hold real value for your organisation?

It is a core business principle that to run a successful company, assets must be managed effectively. However there is an often overlooked asset that, if correctly harnessed, is emerging as a turning point in competitive advantage, particularly since the rise of automation and increasing sophistication of enterprise wide business management systems such as ERP and CRM software. This asset is the core data produced in abundance by every company in the world. 

Corporate data sets are diverse and include customer data, product data, billing data, inventory data and HR data. It can be a significant challenge to maintain all these different data sets often across multiple markets in a manner that allows the company to extract relevant and accurate information as required.

Failure to do so can result at best with time wasted, sorting through duplicate records and at worst with a supply chain of products that are not tracked, are inefficiently purchased and stored in duplicate in different locations, resulting in significant loss of revenue.

If we focus specifically on customer data, the widely quoted shelf life for a customer data store is approximately two years.

This means that if a company has 500,000 accurate customer records in 2006, by 2010 only 125,000 of those customer records will still be of a suitable standard for marketing, customer relations and sales purposes. This is a key principle that applies to all data: it depreciates over time.

Failure to take responsibility for corporate data can result in business-wide problems.

Let's take customer service as a danger area. Any organisation using a call centre must have accurate and reliable information on each customer at the touch of a button. Increasingly the modern consumer expects to make a single call and the operator to know their transaction history, service history and the solution to their query.

Failure to maintain the customer data asset makes life increasingly hard for call centre staff that cannot access a previous log of customer contact, demographic or service information. The result can be fairly similar to forgetting someone’s name, it prompts an immediately negative reaction and can even impact the brand.

A recent survey 'hanging on the telephone', published by the Citizens Advice Bureau, found that four out of ten people were dissatisfied with their call centre experience and that 17 per cent of people found the most annoying aspect to be a poor understanding of their needs.

Other dangers of poor data quality are numerous, for example misdirected marketing costs are associated with poor customer and address data and inefficient supplier negotiations can result from not having a clear view of the supply chain.

The directory assistance number 118118 puts data quality technology to work

As each business is different, so to are the requirements of each data quality implementation. If a data quality technology is to make the standard it must demonstrate flexibility and agility.

During an ongoing project with the UK's largest directory assistance number, 118118, we have been working to ensure 118118's large database of address and contact information is accurate and reliable. A mammoth task considering 118118’s parent company INFONXX manage over 210 million directory listings.

In this instance the task was to remove duplicate listings from the database, therefore ensuring that each customer is provided with a single accurate and consistent contact number.

The technology employed used fuzzy matching capability, which can identify not just exact matches between data records but also close or fuzzy matches such as, South Street Florists and South Road Florists.

The next step to processing this potential duplicate would be for the technology to examine other fields within the data record, such as the post code and telephone number to be sure of a match.

As well as removing duplicates 118118 have actively customized business rules which can standardise directory listings in real-time as their agents enter information into the directory. It was essential for 118118 that they could manipulate these business rules internally via a GUI, as their business requirements changed over time.

118118 have shown the impact that a data quality initiative can have, now standardising over 40, 000 listing records every day and ranking top of recent Ofcom research into directory assistance accuracy, with well over 95 per cent of queries answered successfully.   

Data metrics - You can only manage what you can measure

Being able to produce quantifiable measurement of a company's data asset is an essential aspect of any data quality programme - you can only manage what you can measure. Over recent years business has begun to view data as a corporate asset, implemented initiatives and created data departments.

However if you ask many executives to value their data asset, the response is often a confused one, with executives quoting BI reporting volumes, ROI, TCO, and help desk performance metrics. These are useful tools, but all are designed to evaluate the business and not the data asset.

Metrics provide the opportunity to evaluate current data quality and business practices, and can help to prompt lateral thinking. For example the recent use of metrics to measure data quality lead a yellow pages firm to actually increase the amount of time its data entry staff spent logging each new record.

Previously the perceived value of entering records quickly or 'efficiently' had overridden data quality. Following the use of a metric to value the data asset, the company shifted its focus from speed of entry to recognise the value of the data asset to the company.

Data governance - who's in charge here?

No matter what type of data your company is seeking to manage and maintain, technology alone will not provide the answer. A solid methodology and data governance policy is essential to maintaining consistent data over time.

Traditionally companies place responsibility for data quality tasks with the IT department, which at first seems appropriate. After all the IT department owns the systems that store, transfer and accumulate data.

However the most successful projects are a combination of skills held by IT specialists and business users who work with the data on a daily basis. The coming together of these two fields of expertise has in recent years created a role often referred to as the 'data steward'.

If we assume the example of managing data within a CRM system (customer relationship management system) we can see the need for a data steward. The initial data integration phase from legacy systems involves analysing the existing data sources that will populate the CRM system and monitoring data quality over time is necessary in the long term.

It is essential that the team setting the business rules to govern these activities realise how data should be standardised, validated, de-duplicated and managed.

Or to simplify, 'how this data is to be used by the business'. For example business users would know the most useful format for a customer address to take or how they would wish to view a customer's transaction record.

Reaping the rewards of accurate enterprise data

The good news is that achieving high quality data is not beyond the means of any company.

The key is to treat data as a strategic corporate resource, develop a program for managing data quality with a commitment from the top, and hire, train, or outsource experienced data quality professionals to oversee and carry out the program.

Then, it is critical for organisations to sustain a commitment to managing data quality over time and adjust monitoring and cleansing processes to changes in the business and underlying systems.

Having efficiently managed data will deliver benefits for the entire organisation. A single view of the customer, with an effective understanding of customers' transaction history, service levels can be enhanced and cross-sale opportunities can be exploited.

Cost savings are another major benefit of understanding your customers. LexisNexis for example have saved over £1 million annually from marketing expenses alone by de-duplicating customer records and improving the quality of address data.

Benefits are clear in the present but reliable corporate data can also drive your company in the future and even direct corporate strategy.

Business intelligence systems are an established aspect of strategy formation, with executives seeking the most reliable information possible before making strategic decisions.

Data quality can significantly boost the accuracy of business reporting and provide the information required to outstrip the competition.  

This article first appeared in November 2006 ITNOWextra.