Instantly Improve Conversion Rates & AOV With Collection Sort Logic
This blog post will show you how to dramatically metrics like AOV by up to 30% by implementing collection sort logic.
The best part about this?
It doesn’t require a team of developers & data scientists to get the 80/20 out of results - just a few hours and basic spreadsheet knowledge.
We’ll be focusing on collection sort logic, which is analogous to the way a physical retail store picks out what is on which shelf, and where it’s placed.
Is my company a good fit? Will this really work? Why doesn’t anybody talk about this…
If you have a catalog of 50+ parent SKUs and have never heard of phrases like collection sort, merchandising or sort logic I would be surprised if you didn’t see lift very quickly.
Across hundreds of companies we have considered investing in, and those we control directly, there has always been a correlation between site performance and collection sort.
Running a 4 SKU DTC brand?
You are not going to see nearly as much, if any impact.
With that said, there’s still a lot in here that can be applied across your site from a general merchandising perspective.
As for why nobody talks about collection sort logic…
I think it’s because it’s really only something companies start to tackle once they are into ~8 figures of revenue - a very small audience…
Nerds of nerds of nerds, or something like that?
What we’ll cover!
Why your current approach hurts your AOV and conversion rates, what collection sort is, other applicable places the same logic can be utilized to drive massive wins.
How to tackle collection sort - broken out by available resources and business size.
Build the right solution for your business - accounting for resources available & impact/effort.
What is collection sort? How am I hurting performance?
Collection sort is the order in which you show products to your customers on a product listing page/collections page.
Those with plenty of products benefit most from implementing a systemized approach because putting the right products in-front of the right customer is a lot more challenging when you have 500+ products as opposed to 5.
If you do not have any logic behind the order in which products are shown to customers, then you are making it harder for them to discover your best products. This results in lower gross profit per session.
Many stores default to “best sellers”. This is very, very bad.
Why?
Creates a perpetual cycle where your initial best sellers are always best sellers because they have the best site placements. New products do not get a fair shot.
A product eating up a collection page slot because it sells well might not be generating much gross profit relative to a different product that could be in that slot.
There are many more reasons, but if the first two haven’t convinced you, nothing will. :-)
The logic you use for collection sort can be applied everywhere!
Email, homepage, SMS, product suggestions, and anywhere else that customers are exposed to a “shelf” of products to pick from.
But first…if you’re not a subscriber…
How to design the right solution for your company
To keep things simple and applicable, let’s say our goal is to maximize the amount of gross profit per session.
When designing your collection sort approach, it should:
Provide the highest amount of impact for the least amount of effort.
Be easy to iterate upon, as it’s unlikely v1 is the best core logic.
Match resources available to you.
Fast to launch.
Therefore, my suggestion is that you start with a very simple solution, test it, and then graduate to more resource intensive solutions as you generate results.
Let's jump in!
Which factors matter most, how do we test factors?
Factors and the weight given to them used to determine product rank vary tremendously by business.
Factor examples and explanations - in no particular order.
GP Per Product Impression
Our favorite, and a bit of a pain to implement without slowing site speed significantly. This measures the amount of gross profit generated per product impression. There are a few ways to implement this with Google Analytics (as of GA4, it’s much trickier, I would skip this unless you have a dev team). Revenue per impression is the easier way to tackle this.
Freshness
This is a big consideration for any business with strong repeat traffic. You don’t want your best customers being shown the same products over and over and over. You also need to find a way to give your new products a chance.
Size Coverage
Very relevant in apparel. Best sellers will tank in conversion rate as you lose size coverage. Big problem for footwear specifically. A common scenario is that a product gains traction, is ranked near the top, then sells out of the most popular sizes, and now all that is left are odd sizes and your product page view to conversion rate plummets. Now your top performing ads suddenly look bad.
Variety
If I am an apparel brand sending traffic to a “best sellers” collection, sorting off best selling products on my site is not going to produce optimal results. If you have a top selling shoe in four colors, that now eats up four slots and potential new customers have to scroll even deeper to learn that you also make great jackets.
Lifetime Value.
If your company is heavy on subscriptions, you need to factor in average gross lifetime profits at a SKU level.
Replenishment.
How long does it take for a product to be replenished? How much capital does manufacturing the product tie up, what are the associated costs? This is a big pain point for DTC brands that manufacture in-house with a tiered supply chain.
AOV by First Product Added to Cart
A fun one! Sometimes, especially in niche hobby type brands, we see certain products trigger larger cart values. A product may not cost much, but it could lead to a larger cart because of what people think of and buy next.
These are by no means all of the factors you can use, they are examples for you to understand the concept and perhaps start with.
How do I weigh each factor?
This varies immensely from business to business. We have worked with multiple apparel companies where factor weighting ended up being completely different.
We can inexpensively, and quickly, test to find out which direction we should go in!
If you’re on Shopify, create two collections.
One with the new sort logic you want to try that maybe emphasizes a specific factor, and another baseline collection.
If you’re on Shopify, using a tool like Intelligems, you can set up a simple URL redirect test - no developers needed.
Use the redirect test to split traffic out between collections.
When creating your ranking algorithm (generous use of the word), the easiest approach is to make a G-sheet with all of your SKUs, the relevant factors you want to use to generate your ranking, etc and then upload the final output via a merchandising app on Shopify.
Wow, it worked! Now what?
Your slapped together duct tape solution worked! It looks like the lift justifies a 6-figure expense.
Where do you go?
The “gold” standard is implementing personalization via something like Dynamic Yield. It allows you to personalize experiences to your heart’s content. If you’re in apparel and want to customize your site based on the location of the user to do things like show season-appropriate clothing, you can. The downside is that this will be very expensive (~$60,000 - $120,000) and time consuming to implement.
Custom sort for each ad, especially if it’s already pushing to a collection page. For example, if you’re running a computer retailer and someone clicks on a Google search ad for “gaming PC over $2000” you’d want to serve them a collection presorted for gaming PCs over $2000.
You can use this everywhere. Apply the same merchandising logic to build out your navigation on mobile, desktop, and the footers. Or use it to determine which products are shown to customers in marketing emails.
Did this help? Please send it to a few friends in eCommerce!
If you know anyone in eCommerce/DTC looking to raise capital, send them to us. We are flexible, and invest in turnarounds, growth, and more. Learn more here: www.kartaventures.com
Oh and…not a subscriber? It’s free…