Before you can compare prices with your competitors, you need to make sure that you’re comparing the same or similar products. This is called “product matching” or “product mapping.”
Automated mapping is relatively simple in some categories, such as electronics, where software can easily compare the model numbers on a TV or tablet – although a few retailers may make price matching difficult by stocking lots of exclusive products.
This exercise is far trickier when it comes to clothing or home furnishings, where there are more variations in styles and colors.
Take a look at the different mirrors pictured below. Because the model numbers are not universal, you need to choose the most relevant product characteristics as your basis of comparison.
Which mirror styles should be grouped together as similar products? How thick can the frame be? Which materials? Should a square mirror be compared with a rectangular mirror or are there a significant number of shoppers who are rectangle purists?
Category managers with experience in home décor will know a lot more than a computer about how customers think when comparing similar, but not quite the same, mirrors.
Without any human intervention, the accuracy rate of automated product matching is generally low – dipping below 50% in several categories. With the applied knowledge of a category manager or category researcher, product mapping systems can deliver up to 98% accuracy.
Below is a look at the industry averages for automated mapping accuracy before analysts fine tune the results:
As the expression goes, “Almost only counts in horseshoes and hand grenades.” Even when a retailer hits the 90th percentile, he or she continues to shoot for 100 percent.
As is the challenge with refining any computer search, automated product mapping includes numerous irrelevant and redundant listings that dilute the value of your Pricing Intelligence. A quick search for coffee makers on Amazon produces 31,109 listings alone. Best Buy serves up 1,185 and Walmart has 1,043.
How many do you really need to care about? Before your pricing or assortment analysts determine which suggested matches are ones that matter to your bottom line, the raw data needs some human filtering.
Here are the six steps you should follow to fine tune your product mapping:
- Remove duplicates and irrelevant items from the results stream: Sometimes items are inadvertently labeled with the wrong model number and are miscategorized. A rice maker, for example, may wind up with the blenders.
- Establish which attributes or features are most important: Your search can be narrowed by brand, size, shape, material, color, etc.
- Normalize units of measurement: A King size bed is 76” x 80” and is also known as an Eastern King bed. A California King bed, marketed toward taller people, is 72” x 84”. Make sure your measurements are uniform with your descriptions.
- Identify which private label product features matter most: Tracking down non-branded product matches can be like herding cats. If you are selling refrigerators, choose which features your customers care about most: freezer space, ice makers, slide-out shelves, etc.
- Knock off the accessories: If you search for consumer electronics, your potential matches will have lots of false positives that are battery chargers, protective cases, cords, etc.
- Identify possible overlooked categories: Sometimes your product may be categorized in two different areas. For example, a folding fabric chair might be listed under lawn furniture or beach furniture. A hammock might be with camping gear or with patio furniture.
To further explore how to dramatically improve your product mapping, download a complimentary copy of our new eBook, “Pricing Intelligence 2.0: The Essential Guide to Price Intelligence and Dynamic Pricing,” for retailers and brands.
Get your copy today!
Mihir Kittur is a Co-founder and Chief Innovation Officer at Ugam. He oversees sales, marketing and innovation and works with leading retailers and brands with insights and analytics solutions around their category decisions to improve business performance.