Ensuring product data is consistent across a business may seem a trivial matter. But companies ignore this at their peril, says Aubrey van Aswegen, MD of Knowledge Integration Dynamics (KID).
The way in which product data is recorded in stock management systems may seem an insignificant, local issue. But information from those systems is rolled up into enterprise-wide operating systems, such as ERP, and eventually into management information.
What starts as a simple misspelling can end up with the board seeing double. Product data is difficult to standardise and subject to data quality processes. Few organisations have made an effort to impose an official taxonomy for products. Every time an organisation merges or acquires, the challenge of aligning product data multiplies. And there are no external reference sets against which to validate product information
But taking control of data quality reduces costs, lowers stock levels, improves productivity and provides accurate information for more informed decision-making. To reach this point often means unpicking existing practices and changing how product data is reported and transferred. Even basic productivity applications can add to the problems. If data has a unique identification number that starts with a zero, when this is imported into a spreadsheet, the leading zero will be stripped out. The same is also true for telephone numbers. Basic data quality routines have to take into consideration where the information is likely to be used. If managers need to be able to analyse product data using the application, then product records need to be standardised in a way that will not get corrupted, perhaps by using an alphanumeric format.
Everything is a product
Product data initiatives do not all come from global enterprises - they can be equally effective in smaller companies. The key is to quantify everything and call it a product, so you can see what you are doing for clients in terms of profit and loss. If you are not recording that, people do not record their time properly.
In a service-led business, time is the asset, rather than raw materials. Ensuring that billable time is accurately recorded against client projects delivers vital insight into company performance. Defining the tasks carried out is central to this. Mapping is easy to define; yet it is more difficult with analysis. But you have to give it a definition, and then a product name.
By understanding the amount of time invested against the reward earned, the business can understand its profitability at product level. This is surprisingly rare in the data market.
In many businesses, board-ranking reporting is still taken at an aggregated level.
This allows any problems in the underlying product data to be hidden - or just not to be recognised in the first place. But equally, it means the business does not have a firm grip on some of the key drivers of profitability. It affects everything.
If you had a conditioned view of the product, would you be able to manage more effectively, in terms of what you pay, for example? Yes, because you can reduce costs and increase margins.
It might have been expected that these problems would have been ironed out during the major supply chain management efforts undertaken throughout the '90s. Implementing and integrating multiple operating systems ought to have brought to light the need to consolidate product data.
But most companies have remained focused on the challenges of yesterday, which have principally been dealing with merger and acquisition activity. Simply unifying disparate divisions around one reporting standard has been enough of a challenge. Now they are waking up to master data management as a concern.
Some are looking to hold it centrally under a data steward. The emergence of a new job title is often the indicator of a significant trend. The role of a data steward is to set the product definitions and standards, and then ensure they are actually adopted and used across the business.
You have to propagate a standard master dataset to all points of contact. That solves the problem by presenting a single product data list at the point of input. Customer data has long benefited from look-up tables in this way, allowing the correct address to be returned, even if names are still misspelled. The new architecture of product data involves using Web-oriented service architecture to call up a product dictionary from a central server, rather than having to hold it locally.
What starts out as an apparently simple exercise - trying to align sales data for reporting purposes - can soon escalate into an enterprise-wide challenge involving new processes, functions and technology. But scaling up the issue in this way is also the key to getting significant gains. Every product has a cost, so every piece of product data has a margin.
For more information contact Aubrey van Aswegen, Knowledge Integration Dynamics, +27 (0)11 462 1277, [email protected]
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version