”Once upon a time” when the most complicated engine was Materials Requirements Planning (MRP), and the most common planning tool was MS Excel. Many would say that things were much simpler back then, and they probably were. We were used to stock-outs and lots of stuff on sale because of the misalignment between supply and demand. Oh, yes – the good old days.
Things have changed quite a bit since then. Order online, pick up in-store, or have it delivered to your home. See stores often have the same insight as store associates, to mention a few convenience features that and commonly used by consumers. These advances are advanced engines for things like Forecasting and Replenishment (F&R), analytics, and Machine Learning(ML)/Artificial Intelligence(AI) have been introduced. They all have a common denominator: the determined outcome is less transparent. In the case of ML and AI, it is common that the retailer or fashion company does not at all understand the algorithm behind the results. This means that the associates of the organization are moved further away from understanding how the results came about.
There is a saying,” Garbage in Garbage out.” It has been around as long as I can remember, so there is an irony in the fact that with the implementation of these 21st Century tools client are maintaining items/ Products the same old fashion way directly in the ERP system with no or very little validation of the entered data and without clear responsibility and ownership.
The result is that the outcome of the tool is to various degrees incorrect. At the point where when someone questions the data, the damage has already happened, and it is a very challenging process to correct the data and ‘recalibrate’ the different systems to again get correct results.
Over the years, the realization has changed, partly driven by the realization that the promised ROI did not materialize. The focus on data has increased even though the bigger more significant focus attention many retailers and fashion companies are reluctant on hiring to hire more resources.
The solution has for years been to push the data entry to vendors and other external partners and hence save the data entry and first initial validations. Most retailers have realized that a vendor may not pay as much attention to the data as themselves. The result is data that may not have the expected quality. This has to some degree takes them back to the fact that internally they must maintain data manually.
The good news is that help is here. AI and ML can help to elevate the vision of better data quality with less resource investment. Validating that rich content is aligned with the actual item to make that Taxonomy from various sources is consistent. Have systems suggest validation rules to check data based on existing data and the accumulated mistakes.
On the vendor portal, AI can validate rich content like pictures if the right images are loaded for the correct items and in the correct quality. Something that earlier was labor-intensive. The promise of good quality vendor-provided data may just have arrived. HTS codes (Used for determination for import tariffs) have long been dreaded as they are difficult to determine. An example can be a jacket with a detachable liner. This jacket needs two HTS codes. AI has shown to be able to assign codes with higher accuracy than humans and doing it much faster.
After all some Fairy tales do have happy endings.
To learn more about Utopia's full suite of solutions for the Retail, CPG and Wholesale Distribution Industries, click here.