Success factors part 2
In our February 5 insight we discussed the first three success factors for cleaning product data, which centered around human actions and product information content. In this second insight article, we’ll talk about the last four factors: the process behind product data cleansing.
Our professional opinion is that cleaning product data is not a project but a continuous process. Mistakes will occur wherever people work; “to err is human,” as the saying goes. So where product data is used, quality issues will inevitably arise. Cleaning product data usually starts off as a project, but in the long-term, it should be an integral part of the management–and management process–of product data in your organization. Working with good quality product data not only makes work easier but also more pleasant.
It is crucial to show the people involved in the cleaning process why exactly they are doing the cleaning. The moment people become aware of the importance of their work and the value they bring to the table, they pay more attention to the process. After all, people become more motivated in their jobs when they understand that what they do actually matters.
In order to do this, several options are available. For example, you can organize a review session at least once every four weeks in the cleanup process, where the product data experts show what they have completed. Allow the business side to speak during those sessions, and they can indicate the extra value that the cleaned product data has added to their work. It’s also a good idea to set up a system of Key Performance Indicators in which results can be measured and displayed. Make sure the product data experts can be proud of their work!
As stated previously, when cleaning product data, it’s important to know why this process is being done. Other parts of the product data are important for every business purpose. If your project (or rather, your process) is all about giving the customer a better experience by improving the findability of products in your webshop, you would clean completely different attributes than a situation where your objective is generating short invoice descriptions. Then again, being goal-oriented isn’t the only important aspect; being product-oriented is vital, too. For each product class, customers look at more than one recognizable aspect of the product to make their decisions.
For example: a customer looking for shower faucets usually wants to know if it contains ceramic discs, whether it has a mixer tap, or what material the faucet is made of. But a customer looking for LED lamps would want to know its Kelvin degrees, the number of lumens it has, or perhaps the type of fitting the lamp comes with. So a good first step is to determine the goal and the product class, then see which attributes are important. If you can clean product data quickly, you create added value for the business.
In almost 20 years of experience with cleaning product data, we’ve seen one thing that has proven true time and time again: if you want to do everything in one go, you will no longer see the forest for the trees. Therefore, it’s advisable to not try to clean all data from all products at the same time. Working in small iterations (that is, in incremental steps) in the cleaning process is best. For example, choose a set of product classes that you’re able to tackle in an Agile way in two to four weeks.
First, start with a set of easy classes so you can get a good grip on the process. Then you can quickly scale up to the fast runners in your organization, in order to improve sales and margins. In addition, it’s important to remember that you shouldn’t tackle all the attributes of each product at the same time. Take only those aspects that fit the purpose of that particular iteration. Other product attributes can be dealt with in later stages.
Finally, when cleaning product data, keep the “KISS” principle in mind: Keep It Simple, Stupid! If everyone had their say about the results of a clean-up, you’ll quickly find yourself floundering in the weeds and unable to achieve the results you want in a timely manner. Make one person responsible for the final Quality Assurance step for the cleaning process.
Over the years, we’ve seen projects fail and never become an integral part of the management and control process because too many people thought they needed to give their input. “No, ball bearings are approximately this or that diameter with that particular product code…” or “For fixing this or that, it’s more important that the customer see the color of the screws and not whether it’s stainless steel…” and so on. These are all opinions. Important, perhaps, but taking all opinions into consideration just means immense delays.
Clean product data is a process, not a project. It can start as a project, but ultimately make it part of your process and an integral part of your system. It is a regular, repetitive process that can bring great benefits. Repeat after us: