In digital, you don’t tend to hear the word “merchandising” very often. To me, it’s a phrase that has quite a lot of offline connotations, and whilst most brands will advertise for “online merchandisers” we don’t really talk about it as an area of digital, especially not to the level that we talk about SEO, PPC or CRO.
Merchandising tends to actually fall into the category of Conversion Rate Optimisation, which may be correct as a properly merchandised range will do fantastic things to your conversion rate. But whilst I’ll agree that there are severe overlaps between CRO and merchandising, I find that CRO it is too often confined to site design and content, where merchandising should deal directly with your products and the way that you promote them to users who have reached your site.
Doing this properly utilises so much more than just Conversion Rate Optimisation; this is a bit of CRO mixed in with User Experience, mixed with your Analytics and topped off with sheer industry understanding. I want to look at how some brands have approached this – understanding the way that their customers search and behave to showcase the right product to the right user at the right time. So many eCommerce sites are templated out from an archaic “norm” of how eCommerce is presented, they might look good but the functionality behind it all is how good e-commerce sites succeed.
Know the best way to sell your merchandising products
This is actually marketing 101, right back to the basics! But it goes back to breaking out of the historic ways that eCommerce has typically worked. In order to succeed you need to go beyond what is categories and subcategories to understand how your user wants to buy products and how you can structure your site to make it as easy as possible.
Dunelm were faced with an issue of trying to sell multiple products from a single range on their site. Their users tended not to shop via typical eCommerce categories like pillowcases, duvets, etc. rather they bought across single ranges to create sets for their home.
So Dunelm tackled this by creating range categories that spanned categories, grouping together bedroom items that could all be used as part of the same range.
This range category also includes other neutral items that compliment the range so that a user can kit out their entire bedroom without leaving this page. To make things easier, they have a very easy select system where you pick out the items you want and then add them all to your basket with one click.
Joker’s Masquerade Fancy Dress
This site was faced with a similar story as Dunelm but dealt with it in a slightly different way. Joke knew that users bought by deciding the costume first and then choosing accessories that go with. To make this easier, on each costume product page, Joke created a “complete the look” function that included the best accessories that go with the costume. Users could select the ones they wanted and then one-click them straight into the basket.
Sitting on the other end of the scale, your customers may not be interested in buying multiple products or putting together a range. In the more high-end industries, users are much more interested in spending time comparing products and understanding what the best option is for their particular need.
Prestige Time is an American shop that sells high end (very high end!) watches. One of the core functions of their site is a comparison tool that compares customer inputted products against one another to avoid the user opening multiple tabs or having to switch between products. This easily allows users to make informed purchases without having to check stats on 3rd party sites (with ads to competitors!).
With some extra input, if the user does not fill up all 4 comparison slots then this tool could offer additional suggestions algorithmically based on user search/comparison history.
Don’t Up-sell/Cross-sell/Down-sell Just For the Sake of It
This is a really common mistake for a lot of sites and the effect ranges from not really doing anything to… Well not really doing anything. Default e-commerce practice says that you should make suggestions to users on what they could buy, when they are looking for certain products. What this often results in is unfocused upsells that bare no relation to the user’s behaviour or what they’ve just bought, just what the company wants to sell them.
The opportunity, however, to place other products in front of users is huge! But in order to do this effectively, companies must realise that every single user visiting your site is different and the products that they react will be tailored to them. Algorithmic and context merchandising are really important to suggest the right products to users, based on what they’ve viewed previously or what they are looking for.
John Lewis merchandising
John Lewis, on the whole, have an excellent site with a great user experience, they merchandise their products very well! One of the things they put a lot of effort into is the cross-selling of other products, whilst users are on the product pages.
They have 4 different sections dedicated to cross-sells:
- Customers who viewed this went on to buy – shows similar products that the customer may want to buy instead
- Customers who viewed this also viewed – shows similar ranges to showcase other products that go with this item
- More from this brand – showcases other products within the brand, recognises that customers have brand loyalty
- Your viewed items – quick links back to previously viewed items for users to potentially make comparisons or find where they’ve been
Whilst this is nicely laid out and well labelled, there are 4 sections that pretty much show identical products… Having done a bit of browsing I don’t see the difference between “customers who viewed this went on to buy” and “customers who viewed this also viewed”, which creates a lot of wasted space. When you’re looking at the kind of additional products to merchandise it must go back to what matters to your audience, John Lewis could experiment with a “cheaper alternatives” feed – although they don’t necessarily have a very price-conscious audience. They could also go with combinations or bundle deals that upsell and package add-ons to go with the product, offering a discount. Or they could simply offer recommended products to the user based on the knowledge they have on them; male, big spender, etc.
The kinds of products that you show users as upsells on a product page is completely dependent on your industry and your customer base. Always test different cross-sell phrasing and purposes and experiment using algorithmic product serving to show tailored cross-sells to your users based on:
- price brackets that they’ve been looking in
- colours they’ve viewed
- specific functionality
As another point, what John Lewis don’t do is upselling at the point of sale. Once a user has placed an item in their basket then that’s a pretty strong intent that they may want add-ons, John Lewis should push a product feed into the basket page that shows related products to what’s in the basket, or previously viewed items not related to what’s in the basket, that can upsell to a user just as they’re about to hand over their money.
Categorisation, Filters and Product Refinement
This is a really obvious one but so important to make sure that the right products are pushed in front of the right users. Often so many sites will default into industry standard categories without recognising that users might not want to flip between categories but rather filter down according to their preferences.
Asos are always a good example to use for digital best practice, their categorisation and filtration function is second to none and a fantastic example to use for showcasing huge ranges!
Within the male and female categories Asos categorise as you might expect from a clothing store however they still have often nearly 1,000 products per category. This many products is typically an e-commerce nightmare, however the use of filters to narrow down across colour, brand, price or style allows you to wade through everything you don’t want to create your own category of just the products that you may consider. Is this a better way than fixed categories chosen by the site? Yeah, I think it is!
This way of browsing is fast becoming normal for e-commerce online, where Asos stand out is that you can easily create combinations of filters to include all the products that you want without having to move between categories or filters. Consider the way your categories work and how you can move away from fixed business formed categories, and instead serve the user with the exact selection of products that they want.
These are some great ideas as to how you can approach and potentially change the way that you merchandise on your site. Underpinning them however are two factors that are crucial to businesses being able to do replicate and merchandise properly:
Tagging your products
A lot of what I’ve talked about above is done algorithmically, what’s not done algorithmically is done manually and having your products properly tagged makes that much easier as well! This is a very long and drawn out process of attaching tags to your products that help identify and organise them to help your software know which ones to show when.
Whilst this is a painful task, you can take solace in reading one of my favourite ever articles – it talks about how Netflix created their movie genre generator by tagging Hollywood through watching and categorising each and every movie in their database. This approach gave them a famously powerful genre generator that is essentially a merchandising powerhouse! By taking a similar approach, you can pivot your product range to the nth degree, offering precisely targeted products based on what the user wants.
Data underpins all of this, whilst a lot of it comes from industry intuition, it is so very incomplete without rigorous testing and measurement. How do you know if your upselling feeds are performing better or worse as you change the phrasing? You can measure the clicks on the feed products and the conversion rates of users who used them. How do you know which items can be paired with others? Transaction data.
Most merchandising practices are interaction based, which makes it difficult to track via regular click streams. If you’re using Google Analytics then make sure you’ve got some event tracking on your key merchandising areas and monitor the effectiveness of the algorithms that you’re using.