General information about PriceSpy indexing process
Here at priceSpy we handle an incredible amount of data. One of the largest areas for us is our indexing, the process of how PriceSpy connects a shops offers to our product listings.
This happens either with manual indexing or through our automated system.
The way our manual indexing process works is that our colleagues manually map the data we get from shops through feeds or scanners and then sort that data to fit the right product to the correct product listing.
Does the shop have a well structured product name this process becomes a lot more effective and simple. With the automated indexing PriceSpy uses primarily different types of codes, that means it is very important for shops to supply us with the codes either on your website and/or in your price file.
When there is a new product release or when PriceSpy does not have a product listing for the products a manual touch is required, we are not able to show a price comparison until a proper product listing have been created. If a shop has a code connected to its offer we will be able to read that code and connect it to that product listing, and that opens the possibility for the use of the automated system for shops with the same codes.
In general the automated sorting is more effective and faster than what the manual indexing is.
A little more in-depth
PriceSpy indexing happens in 5 different stages.
A shop registers on PriceSpy - this happens one of two ways, a shop notifies us of their interesting in being on the site or we find a shop and adds it.
We create an agent reader for the shop. First we ask if the shop is able to supply us with a price file, if the store does not have a price file we create a scanner to scrape their inventory instead
Category sorting - This is the first step when we have read in the shops information. We map the shops categories against the categories we have on our website. The products we do not have an active category for does not gets sorted, that means PriceSpy excludes the category for sorting. Excluded categories will show up on our website via PriceSpy RAW. The reason for this step is because a shops offers comes as big data packets, by that we mean that the offers can come in packets of hundreds (sometimes thousands) of products and this step helps immensely with the continued sorting for us.
Product/Offer indexing - After the category sorting the shops offer will show up in our system so we can handle every product individually. If the shops has a code connected to the product ( and we have that code connected in our system from earlier) it will be sorted automatically, otherwise we would have to do a manual indexing.
To product - When the product is in the right place and we have indexed one or more offers in our system to it, it will show up on PriceSpy. Our users can now find and compare prices for the product.
How does manual indexing work?
When we manual index we use the information supplied to us by the shops.
Primarily we go by the product name and what the shops call them.
A well structured name includes manufacturer and model code / name. In some categories it is also important to supply us with the size of the product, for example shops that sells perfumes. Colour and materials is not always necessary to include but can help with our indexing nonetheless. We also prefer that there is a manufacturer code included on the product in the feed or on your product page. This makes it easier for us to differentiate between similar products.
When PriceSpy mentions SKU we mean the manufacturer code, MPN.
Product name examples
A good product name - no unnecessary information
PriceSpy reads the model code/manufacturer code (MPN) from a SKU column
The shop also supplies an EAN-code, that is read in its own column
Contains manufacturer and model name, the shop does not add any extra unnecessary information to the name
2. An alright product name - Important information mixed with less important information
Code is present
Model and manufacturer is supplied
A lot of unnecessary overflow of information that is not needed.
3. Not a good product name
A very general name for the product
Missing model code
SKU/MPN- and EAN-code are missing
When a product name is to general and we can't for sure tell which version it is the product doesn't get sorted. ALL products from shops are searchable via our search function. If the offer is not on a product listing it will be found in PriceSpy RAW.
What is important for the success of the indexing process?
A well structured product name with connected codes makes it the easiest for us to index your offers.
- Official product codes, this can be an EAN, Manufacturer Part Number (MPN) or an SKU-code.
- Article numbers or unique codes that is only for one shop does not help our indexing process, as these codes normally is not bound to a product thus pur system won't recognise the code and will not be able to use it.
Good product names -
- We would like to see the product names as clean as possible. We ask you to exclude longer product descriptions unless they are necessary. A good product name contains manufacturer, model name/model code as well as the size/volume (if applicable). Colour and/or materials are not necessary but typically they help our indexing.
- Size/volume - In some categories it is very important to add sizes and/or volume. Some examples are Perfume, Watches etc.
- Version - Some products have different versions. This could be a special release, a new generation of the same product or a new release with a similar name.