Manufacturers and distributors want to know how their products compare with the competition, in price and in specifications. The data usually exists publicly: online shops, price lists, technical data sheets, manufacturer websites. We can collect it automatically and regularly. Most of the work, however, is not in collecting. It is in matching: deciding which competitor product is truly comparable with yours.
I use the same approach in my own project WatchBI, which collects prices and specifications of new luxury watches from retailers in Europe, North America and Asia every day.
Collecting the data
The sources vary: product pages, PDF price lists, downloadable data sheets, online shops. We take as much as possible from them – not only price and name, but also technical parameters, usage instructions, images and documents. The more we know about a product, the more accurate the matching.
Collection runs automatically once a day, early in the morning, so the data is ready for the start of the working day. A new document, such as a data sheet, can also be processed right after upload.
Matching: you have to understand the product
Comparing two products by name is not enough. The same item has a different name in the price list than on the website, often in another language, and names can be four or five words long. Matching therefore works with the full picture of the product: name, description, use and technical parameters. Only then can you tell whether two items are comparable.
Watches are similar. The same reference is named differently by different retailers, and whether it is the same model depends on the parameters: case diameter and material, movement, dial.
Comparing per unit
A price per pack says nothing if one manufacturer sells 5 kg and another 7 kg. The comparison therefore converts prices to a unit – kilogram, litre, square metre – and for watches to a single currency. Only then does it make sense to say who is more expensive.
An expert confirms the match
Automatic matching proposes candidates and a confidence level. Whether two products are truly interchangeable is a decision for someone from the field. I can build the system, but I am not an expert in every industry it runs in. The application therefore has a simple interface where an expert confirms or corrects the proposals.
A source for every data point
Every number and every image should have a traceable origin. On WatchBI the source is shown right below the gallery, and each model has an overview of all images with links to where they come from. In a company this matters even more: when a salesperson shows a comparison to a customer, they need to be able to say where the price comes from and from which day.
What the comparison is used for
- Pricing – where you are well above or below the market.
- Sales materials – comparison tables for a specific product and competitor.
- Assortment gaps – where competitors offer a product and you don't.
- Quotes – up-to-date data for price quotes.
The results can be shown in dashboards with a chat over the data, which I describe in Custom reporting with a chat over your data.
We build custom market and competitor monitoring platforms at OREGO Productions.
