WooCommerce is an exceptionally powerful, customisable retail CMS plugin for WordPress.
As it’s WordPress, some organisations tend to reject it in favour of more complex platforms like Magento. While that’s probably a better choice for larger enterprise retailers, you’d be surprised by how much WooCommerce has to offer, especially for start-up retailers.
With that said, Woo naturally has some limitations. One of them is the native related products functionality. The default “Related” product recommendations work by tag or category and are a function of a simple match. You can exert more control by creating “Upsells” but these require manual configuration in the product CMS.
While Upsells are a really nice way to make better product recommendations, setting them up can take a long time for a site with 2,000 products!
Despite WordPress having lots of related posts plugins available, I’m still yet to find a related products plugin for Woo that works the way I want it, so I came up with a process to create a related products table that you can upload directly into WooCommerce’s upsells feature.
Of course, this is a process that can be modified to your own tastes. You need to be able to export and import data into WooCommerce (we use WP All Export and WP All Import) and you need to have the Fuzzy Lookup Add-In for Excel installed.
With this process you could:
- Make better recommendations based on product views and purchases
- Perform product matches by attributes not featured in Tags or Category names
- Avoid the use of post tags altogether
- Combine external data sources, like SEMrush to order matches by search demand (see a post on a related topic here)
Put simply, you can build your own product recommendations model using all sorts of attributes normally reserved for very powerful, enterprise systems.
Minor caveat: as with all home brewed methodologies the results aren’t always perfect and it’s, therefore, wise to have a few iterations until you’re happy with the results. Also, it’s easy to keep on plodding along for too long, with the spectre of diminishing returns fatiguing your efforts.
For me, as long as the product recommendations kind of make sense and you’re saving time in the process, that’s a win. I particularly like anything that improves my internal links in sync with search volumes. Have a think about ideas like that while you’re soaking up this methodology:
Here’s how I create a related products table that is importable into WooCommerce.
Export the data with WP All Export
First, we need some product data to work with. Head to WP All Export and select Specific Post Type. In the drop-down, find “WooCommerce Products”. Then in the options below, click the Customise Export button.
When you customise your export, you perhaps don’t need every attribute in the CSV. I have a preset export template that looks a little like this:
The main data points I tend to always use are short, text-based descriptors such as title, tags, colour and category.
By the way, don’t be greedy. If you use the actual product description for the fuzzy matching process below, you’ll quickly run out of power. There are too many characters for Fuzzy match to be able to handle.
Download your data and open Excel.
Processing the data in Excel with Fuzzy Lookup
Firstly, create an Excel file. Mine’s called “Playground with transpose” it’s an exciting and appropriate name because I’m going to be creating a macro to do a bit of clever transposing later on.
For now, here are the important details:
I’ve created two identical data tables on Worksheet A and Worksheet B (highlighted in blue), sorted in opposite order by ID value (highlighted in Green).
Why two tables? Fuzzy match works from two tables and in a linear fashion, so I believe that having two tables in reverse order of each other will allow the tool to deliver a good breadth of matches. I could be wrong about the linearity; it’s just a hunch but seems to work well. You could also sort by page views or order volume as a measure of popularity. This is where data from external sources could add an awful lot of value to the results!
I’ve named each table very imaginatively as TableA and TableB (highlighted in red).
Next, I’m going to head to the Fuzzy Lookup tab. Make sure it’s installed, create a new tab (I’ve called mine “Fuzzy”) and click the Fuzzy Toolbar button.
You need to think about the data points you match on and in what order. This is very definitely a place for trial and error.
In my example, I’m going to match by Product Category, then Brand for simplicity. You could match by Colour and Product Title (or lots of unique combinations). It’s just something that would be unique to the type of products you have.
I deselect most output columns for my production data as you’ll end up with duplicates – but make sure you’re keeping what you need and don’t be scared to throw away the results and try something new.
Save your spreadsheet! When you click “Go” you’ll immediately wish you had as Excel will look like it’s crashed and your laptop cooling fan will spin up:
After a bit of time, you end up with a result. I have to admit – 12 related matches on two match columns was too much for my little i7 processor to handle, so I switched down to 8.
Naturally, it’s important to check you’re happy with the outcome and re-run the process if not. There are lots of different ways you can match products to similar products and lots of ways to sort the match order (sales volume, view count and so on), so don’t just settle for the first result!
Prepare the data for upload
For this next step, you need to copy the product ID list from the left and the product ID list from the right into a new table:
It’ll look something like what’s above, only longer.
Once you’ve created your table, you’ll see the problem. It’s a 2 column table with many duplicates in the left-hand column and some duplicates in the right. Duplicates between the left and right columns would mean your related products would recommend themselves, so best to clean that up. Managing the duplicates in this way is easy, you can just use
=EXACT() to find them and delete.
Transpose the data into columns
When you’re done, we still have the problem of many duplicate IDs in the left hand column, so we need to transpose the duplicate rows in to columns in Excel. This handy VBA Macro from KuTools is an absolute lifesaver:
Dim xArr1 As Variant
Dim xArr2 As Variant
Dim InputRng As Range, OutRng As Range
Dim xRows As Long
xTitleId = "KutoolsforExcel"
Set InputRng = Application.Selection
Set InputRng = Application.InputBox("Range :", xTitleId, InputRng.Address, Type:=8)
Set OutRng = Application.InputBox("Out put to (single cell):", xTitleId, Type:=8)
Set OutRng = OutRng.Range("A1")
xArr1 = InputRng.Value
t = UBound(xArr1, 2): xRows = 1
.CompareMode = 1
For i = 2 To UBound(xArr1, 1)
If Not .exists(xArr1(i, 1)) Then
xRows = xRows + 1: .Item(xArr1(i, 1)) = VBA.Array(xRows, t)
For ii = 1 To t
xArr1(xRows, ii) = xArr1(i, ii)
xArr2 = .Item(xArr1(i, 1))
If UBound(xArr1, 2) < xArr2(1) + t - 1 Then
ReDim Preserve xArr1(1 To UBound(xArr1, 1), 1 To xArr2(1) + t - 1)
For ii = 2 To t
xArr1(1, xArr2(1) + ii - 1) = xArr1(1, ii)
For ii = 2 To t
xArr1(xArr2(0), xArr2(1) + ii - 1) = xArr1(i, ii)
xArr2(1) = xArr2(1) + t - 1: .Item(xArr1(i, 1)) = xArr2
OutRng.Resize(xRows, UBound(xArr1, 2)).Value = xArr1
To create a Macro, copy this code and head to View > Macros. Give it a name, “TransposeUnique” and click “Create”. Paste the code into the Macro editor and save it.
Then, press the F5 key to run the Macro or head back to View > Macros, select your new Macro and click “Run”.
Once the Macro has run, your table is ready! I copy the results to a new tab and then save that as a csv file:
Upload into WooCommerce
I mentioned we use WP All Export for the exporting process and funnily enough, we use WP All Import for the import process! I created a step by step to upload your CSV file once it’s ready. Happy uploading!
- Upload your CSV Data:
2. Check that the import data looks right before you proceed:
3. Map data to import with the WooCommerce add-on:
4. Being sure to add all variables for each ID:
5. Match your products with cross-sell products (using the parent ID in this case):
6. Choose which product data to update (only cross-sell data):
From there, run the update and let the process complete before you check the pages are updated.
As I’ve mentioned several times, I’d strongly recommend you try a few iterations of this process to suit your own needs. This article is very much a prototype – in fact, the process changed for me while I was writing this article!
I do think, though, that there’s an opportunity to include data from external sources to drastically improve your related product targeting, particularly from an SEO standpoint. Have fun experimenting with that!