CPG manufacturers are taking advantage of retailer data to optimize their businesses, driving sales while reducing costs. This retailer data is being used in daily sales reporting, replenishment decisions, demand planning, data visualization tools, data lakes, and other strategic projects.
The number one concern we’re hearing from the industry is that business users struggle to trust the numbers, because numbers “seem off”, don’t tie between systems or with buyer views, or are just wrong.
Here’s how we are addressing this data integrity problem, and the steps we are taking to ensure that customers can trust the data in our platform.
Retailer data sources can be complex, and offer many choices for very similar data. Make the wrong choice and you won’t be looking at what you think you are. For example, with dotcom data, the difference between ordered and fulfilled is night and day – use both and you’ll double count, pick the wrong one and you could be making a decision based on where products were ordered rather than fulfilled (which can be significant if you are a replenishment team).
Our Customer Success Analysts have intimate knowledge of each retailers’ options, and will guide you through these important decisions. That way, you’ll know how to get accurate brick and mortar numbers versus dotcom channels, and what inventory buckets to select so that you don’t double count units so users will trust your supply insights.
Of course, all users want information ASAP, but this can be tricky because retailers provide data updates on different (and inconsistent) schedules. If you’re looking for sales updates for last week, this data could come in any day Sunday through Wednesday depending on retailer. If you’re looking for daily sales performance it can get even wilder, as some retailers provide data for yesterday, while some retailers can only give you POS for two days ago.
To help with this complex scheduling, our retrieval process that knows when retailers update their different data sources, and pulls the data on the right schedule. It will also automatically reschedule individual data pulls if the data isn’t ready on schedule. Considering that some customers pull hundreds or thousands of data files some days, this “self-healing” approach is critical to making the process work efficiently.
Once you have data, a common mistake is to immediately start using it. The problem here is that sometimes the data is just wrong. You might be wondering how the data can be wrong – the list is long, but it includes missing stores, items and metrics, as well as numbers that are just not within reason. Some customers try to do a quick visual review of numbers each day, but that is a massive and time intensive task, and it delays the entire process.
Our platform performs a 15-point assessment of data quality, accuracy, and reasonableness before data is released for consumption. These integrity checks range from the basics of making sure there are no gaps in the data and no key metrics are missing or zeroed, to more advanced checks that compare new data against historical trends. There are some complex methods in there; we won’t bore you with a discussion around standard deviation and deviation from the mean. If any of the automated data checks are flagged, our platform goes into motion to re-retrieve data to fix the issues so that “bad” data doesn’t creep into the reports.
A key to holding trust with the users is letting them know where things stand. That means keeping your team informed when retailer portals have issues, or if data is held for data integrity issues. It seems simple, but letting customers see the status of their data is vital.
Each retailer has their own view of your business, yet you need to see sales and inventory performance across your retailer channels. Unfortunately, that just doesn’t work without some extra effort.
TR3’s platform includes an on-going process that maps items across retailers, transforming the numbers into customer-preferred categories and views. This isn’t a just one-and-done thing, it needs to be an evergreen process to make sure that new or discontinued items don’t cause issues. That’s why we consider our automated process to be a watchdog, keeping an eye out in the background for issues.
We’ve optimized our processes so that all of this gets done without slowing down the delivery of reports to users. Our customers have the data they need, and trust in it, when they hit their desks each day.