Following his breakout session Preparing for PIM: How to approach an Enterprise PIM Project at the 2019 Akeneo PIM Summit, Akeneo functional consultant Adam Beatty provides further insight into the importance of data governance in a PIM project.
In the digital age, customers crave information. The constant connectivity offered by the internet and mobile devices gives buyers an endless opportunity to educate themselves and scrutinize every facet of a product or service before they make a purchase.
This new era of information means that brands and retailers are finding their products are under the microscope of consumers 24 hours a day. But, what happens if that information is wrong?
Providing customers with accurate and compelling information is the most fundamental purpose of Product Information Management software. The data that makes up this information is what informs your customers about your product, and has a significant impact on their decision to purchase. If you provide them with incomplete or inaccurate information it can heavily weigh this decision against you and result in a lost sale. However, if you are able to provide them with a product experience that’s error-free, comprehensive, and compelling, you can eliminate these lost sales, differentiate yourself from the competition, and even secure additional sales.
Reaching this next level product experience requires a foundation of high-quality data, which itself requires a proper data governance policy. Today we’ll explore how you can provide your customers with this high-quality data as part of your PIM implementation.
Correcting poor data and establishing governance can be an overwhelming task. It’s not uncommon for enterprises to have significant issues with their data for a variety of reasons. This could include multiple acquisitions or divestitures, application upgrades, rapid expansion, or a general lack of emphasis on data governance around their product information.
Fortunately, PIM is here to help.
A PIM project provides the ideal opportunity to implement data governance and cleanse your data.
During a PIM implementation, the cross-functional teams and technical resources needed to fix your data issues are already invested in the project. It’s just a matter of making data governance a priority.
Admittedly, this is often easier said than done, as PIM projects are often implemented as part of a larger eCommerce project, in which time and resource restrictions are a factor. A common scenario we see in Akeneo projects is the integrator or customer deciding early on to just move their ERP data directly into the PIM, without any cleansing effort. This is usually because the integrator or customer is more concerned about tangible deliverables related to connectors, customizations, or migration into the eCommerce platform, and cleansing, unfortunately, becomes an obvious place to cut corners. This is a mistake and means that while your product team may be using a fancy new tool, customers will still be seeing the same inaccurate data they saw before the project.
So, as you implement your PIM project, how can you establish proper data governance? Below are some key steps that must be considered in any PIM project to make sure you’re providing your customers with the best product information possible:
1. Identify the missing or incorrect data
Before you can improve your data you must first identify what information is missing or incorrect. This may seem easy and obvious, but can be more difficult than it appears, particularly for larger catalogs with millions of lines of product information data. Product marketers — or as we call them at Akeneo, “Julias” — are often the most valuable in uncovering bad data, as they have the most intimate knowledge of each product and product attribute. Your Julias can provide a fantastic starting point that your technical resources can then supplement. For example, Julia may know that there are multiple variations of a dimension attribute (3X5, 3 by 5, three by five, etc..). She can then inform a developer on the IT/Integrator team, who can quickly export the data and run queries to isolate the data and see all the possible variations.
2. Establish a governance policy
Once your data issues have been uncovered, you must determine what the correct data values are and establish policies to ensure consistency across your catalog. For example, in the dimension example provided above, the product team may sit down and collectively decide that they want the dimension value to be standardized as “3X5” across the entire catalog. Another example could be more focused on the business process itself —if Supplier A provides excellent product images while Supplier B provides poor low-resolution images, a policy could be enacted to retake all of Supplier B’s images, while automatically using all Supplier A images. Establishing these policies is vital to standardizing your data, and Akeneo PIM provides excellent tools to help with this process. The use of completeness and attribute options are particularly powerful tools in Akeneo that can be of great assistance in building these policies.
3. Fix the source of the problem
Once poor data has been identified, it’s time to fix the source of the problem. The product ecosystem can be complex, and fixing the source of the problem requires you to understand exactly where each piece of product information is coming from. There can be many different causes of the problem including poor supplier information, incorrectly mapped data coming from another system like an ERP, or even human error, like the inconsistent use of an attribute value across a product team. Resolving the issues will depend on the problem itself. A mapping issue will be resolved by the IT team, human error must be fixed by standardizing data entry processes, and external issues, like poor supplier information, could require escalation by the purchasing team.
4. Correct your existing data
After the source of the problem has been resolved (and you’re confident it will no longer create issues) you must update your existing catalog. A data migration (like when implementing Akeneo) provides the ideal opportunity to fix this data. Once your business team has identified all the data issues and determined the correct values, the IT/integrator team can programmatically correct large amounts of data before migrating into the new PIM.
5. Maintain and update your data governance
Product information is dynamic. New products and technologies are constantly being phased in and out, and your governance policy must be designed to accommodate every change. When you add a new product line you must review the new line’s product information and adapt your existing policies and procedures. This is necessary to ensure that you are continuing to only provide your customers with high-quality information. It’s also important to remove data that has become irrelevant. For example, if a brand no longer exists in your catalog we recommend you delete it from your systems. A large B2B or B2C enterprise selling hundreds of brands at a time could easily end up having dozens of outdated “ghost” brands in their systems after several years if they’re not actively cleansing them from their product catalog.
Clean and complete
Akeneo provides many features that help you cleanse product data and establish governance. The use of attribute options, metrics, prices, and custom and reference entities ensure that only correctly formatted and predetermined values make it to your customers’ eyes.
Completeness is another important Akeneo feature designed to ensure a product isn’t exported to your channels until all important product attributes are filled out. Completeness also helps the business team structure their workflow across channels and locales and is the basis of productivity tools like Projects and Views. The rules engines allow the automatic transformation of data by the PIM based on predetermined business rules, and Proposals are a validation workflow to ensure nothing falls through the cracks.
Finally, Akeneo has recently released two new tools, the Akeneo Onboarder and Ask Franklin. The Onboarder allows suppliers to interface directly with the retailer’s Akeneo environment and provide data in a predetermined format. Franklin, meanwhile, utilizes AI and machine learning to provide technical product information about your products, and can also directly interface with Akeneo.
There are also options outside of Akeneo that can be used in conjunction with the PIM to help ensure data quality. Developers can write scripts and utilize regular expressions in their language of choice (PHP, Java, Ruby, Python, etc.) to parse relevant data and cleanse it based on pre-established business logic. Customers can also use more traditional means to fix their data issues, such as emailing or calling a supplier or other internal teams directly and requesting better data.
Iterate to Great
It’s important to remember that it may be unrealistic or even impossible to correct all of this information in a single effort. Instead, it’s best to take an iterative approach. A PIM implementation is a great opportunity to start the process; during this migration, you can focus on the most critical data issues and establish governance using Akeneo’s many available features in the PIM. Once live, you can then progressively work towards improving your data by assigning your product team small side projects around certain attributes or parts of the product catalog.
It’s easy to get caught up on the latest new technologies surrounding the ecommerce space, but it’s important to remember that these new technologies are ultimately just tools. A new tool may dazzle your customers with how it presents your products, but if there’s no effective data governance policy in place your customers will still be seeing the same inaccurate or missing information that could impact their purchase decision.