Installed Base Data Quality: The Achilles Heel of IoT, eCommerce, and Servitization initiatives
As the Covid crisis engulfed the world last March, companies large and small grappled with the existential threats to their businesses. Different sectors were impacted to varying degrees, and it was clear to everyone that the crisis was going to cause habits to change, some irreversibly.
In the Industrial OEM world, we started seeing companies make investments cautiously in areas that could help them with “remote and distance” service to their customers. The belief was, the changes triggered by Covid would stick, and they were willing to invest in small experiments to determine the feasibility, ROI, etc., of these new investments. Specifically, OEMs launched initiatives in IoT, eCommerce, and AR/VR as a way to deliver a high level of service to their customers without compromising the health and safety of their employees.
As the year progressed, a few things became abundantly clear:
- Customer expectations were permanently changed – in a way that challenged OEMs
- IT departments were stretched thin – too many initiatives, too few people
- Initiatives that delivered operating efficiencies were higher in priority and got funded
- Pilots got delayed and slowed down due to a combination of 2 & 3
- A growing realization that data quality and completeness was blocking or slowing these initiatives
Quality comes last
The issue of data quality has hamstrung most companies from making progress, especially as eCommerce, IoT, and Servitization require a single source of customer truth as a necessary requirement to operationalize these initiatives. The range of data quality issues can be as simple as duplicates (or many more!) of customer names, addresses, contacts, etc; misspellings, incompletes, and other data issues. But as the complexity and completeness of the customer engagement increases, data requirements (and therefore quality issues) increase dramatically.
At Entytle, we’ve seen our share of data issues. A memorable example is finding 62 different entries in various systems (ERP, CRM, Service, Support, etc.) for one single customer! Another is large-scale incompletes (e.g., Acme Inc, United States as the complete address from an ERP system). Or worse, unclassified and uncategorized (no taxonomy?) for equipment, parts, assemblies, kits, consumables, services, etc.
The lack of data quality leads to issues with the integrity of the analytics performed with IoT sensor data, eCommerce recommendations, and, more importantly, customer histories. Imagine, for example, with 62 different IDs for the same customer at the same location; clearly, transactions were being logged against each of these customer IDs. So which one should a sales or service rep believe? How does she know what the customer has truly done? Therefore, can the analytics and recommendations have any integrity? When moving to an eCommerce environment, which of those 62 customer profiles/histories should be used to present options to the customer? Without solving this foundational challenge, manufacturers risk taking the investments they are making to create value for their customers (i.e., IoT, eCommerce, Servitization, etc.) and turning that investment into a net negative experience for their customers.
These are just a few examples of what could go wrong with poor data quality. In our next post, we will outline how you can improve data quality without investing years and significant expense in full-scale Master Data Management (MDM) solutions.