DEEPER DIVES: Wardrobing, Waste, and Weak Profits

Good Morning,

First, as always, thank you for joining.

Half way through Q1 and certain trends are really sticking.

Revenue is top of mind. With all of the competition in the market, and a massive fear and uncertainty around global commerce, EVERYONE is looking for more ways to stand out.

From brands to service providers, everyone knows that consolidation needs to happen. Retailers are pulling back to a smaller core contingent of service providers.

Service providers are scrambling to be able to sell on more than low cost.

Here’s what this issue brings:

• While AI is all the rage and people doubled down hard at Manifest, there’s one theme that’s emerging more than anything else - VALUE. All the flash and slick UIs aren’t worth anything if you can’t show tangible ways the tool improves operations.
  
  Here’s another one of my favourite uses for AI in logistics - Return processing. I’m talking here about vision tech, what you need to do to set it up correctly and why exactly it will replace the majority of physical classification and grading. If you are a retailer using a service provider that isn’t using these tools, you’re going to be way behind the leaders in this space.
  

• JD.com is getting into the food delivery business. That’s (kind of) interesting. But the real story is about how they follow the market but most services providers don’t want to do the same
  

• Get yourself set up to be able to test anyone’s solution. Data masking to get an offer from anyone.

Retail’s $101B Fraud Problem Meets Its Match: How Vision AI is Revolutionizing Returns

Everyone needs to sell more.

I posted about the GXO results over the weekend. Despite having a strong end to their year (and a strong year overall), the stock still tanked.

It was all because of the fact they changed their guidance for the upcoming year based on known customer changes.

Because they were going to sell less (at least initially), people started selling off the stock looking for greener pastures.

The pressure to sell isn’t only on service providers.

Every retailer and D2C brand is feeling the pressure to get more people to buy more stuff.

And because of that, returns are at an all time high.

Depending on the segment your products belong to, retail returns will be anywhere from 10% to 35% of your sales.

With the current levels of eCommerce, this works out to about 4 billion parcels shipped back annually.

To keep sales as high as possible, and not lose money on the cost of the inventory, retailers are doing everything they can to be able to resell the products for full value.

This has resulted in returns handling that is costly and inefficient. Recent statistics show an average cost of $20 per return due to manual processing and fragmented logistics.

Completely unsustainable.

The problem is now too big, and brands are giving up on trying to outsell sloppy.

You know from following what I share that I take a cautious view when it comes to AI.

I 100% believe that it will fundamentally change the way we work.

But I also believe that they is way too much hype around some claims that are being made and I also think that it has the potential to create wild mediocrity when it comes to people’s innovation and desire to solve problems in new and unique ways.

(I saw that because I find too many people look at AI as an answer machine rather than a tool that allows them to go deeper).

But, there are some areas where I think AI will be dramatically better than human ownership.

I’ve shared a few areas in older editions, and today I want to talk about another one.

Returns Processing X “Vision” Based AI

The reason that AI will dominate this space is because of one thing above all else.

Consistency.

Humans, unfortunately, are not as consistent as we like to believe.

Different factors can impact our mood and perception, which means that how we make decisions (even with the same inputs) can be different.

And in a world that’s now generating billions of dollars in returns, inconsistencies and altered decisions can add up to a lot of money (not to mention the potential ‘risk’ to product quality).

AI image processing, particularly using techniques like Deep Learning (a subset of AI), has shown significant promise when it comes to quality, execution time and accuracy.

AI algorithms can be trained to detect very subtle patterns and anomalies, much more subtle than would be realistic to assume for a human to perform consistently.

The training of these models also reduces errors of false positives and false negatives. This means that retailers get a better grading score for their goods which directly impacts the customer experience of the resale of that product (whether it’s sold as factory fresh or as a known ‘open box’ item).

AI for returns processing HAS to be the future.

AI will streamlining returns by automating decision-making, optimizing routing, and assisting the overall warehouse operations.

It will reduce processing time and lower handling costs by at least 20%.

It can provide expanded “processing capability” by analyzing customer-uploaded images/videos to determine if a returned product matches the original purchase (e.g., size, color, wear and tear), reducing fraud losses (which totaled $101 billion in 2023).

This option to pre-process or edge process could allow for dramatically different value chains to be created. For retailers with retail locations, they would be able to effectively manage return and resale activities across an entire network, even though each store is staffed with completely different people.

The best part of this system however, is that it also collects tons of data that can be used to improve revenue quality and REDUCE returns on the front end.

AI processed returns that leave the retailer with high confidence allow for dynamic pricing models for returned goods. This ensures optimal resale value by adjusting prices based on demand, product condition, and seasonality.

It also allows for condition-based resale strategies (20–30% faster restocking cycles) minimize warehouse congestion. There are too many brands out there that are holding onto WILD amounts of stock. Those in the apparel and footwear space feel it the most, because of the number of variants that these segments create. This slow cycling out costs retailers millions of dollars every quarter and create increased warehouse costs that only get worse over time when you look at the revenue : cost ratio that develops.

On the elimination side, virtual try-ons and 3D product visualization improve online shopping accuracy, helping customers make better purchasing decisions.

I recently bought glasses for one of my daughter’s and the virtual try on was a massive improvement and made buying something people typically want to see and touch much easier and more comfortable from the comfort of the couch.

Here is my overall guidance for the brands out there.

DO NOT invest in ANY returns solution that doesn’t have these types of functionalities at their core (or currently under real development that can be quasi-demo’ed).

No matter how good the UI is, or how fast someone can be trained on it, you are wasting your time building with any system that won’t allow you to have access to what returns will be like (very soon).

When Giants Like JD.com Leap, Why Are You Still Crawling?

I speak to a lot of service providers.

If there is one thing that I think I can share when it comes to a challenge most of them face it’s on diversification.

The reality is this. No matter what you are telling yourself, you are likely too unwilling to push outside of your comfort zone.

When you think of the type of customer your want, the type of way you want to work and how you are willing to generate your revenue - more often than not it’s a simple extension of what you have already been doing.

To quote Brittain Ladd, there’s a huge gap when it comes to “Thinking Big”.

That’s why I was surprised to see a move by JD.com that was announced last week.

JD has decided to dive into China's $156 billion food delivery market.

In the past, they had deliberately avoided competing in the budget segment dominated by Meituan and Ele.me (which collectively control 98% of China’s food delivery market).

But JD has been struggling. Amid an economic slowdown and weakening consumer spending power, they have had to resort to selling on discount campaigns to counter the competition. These price wars however, have resulted in a decline in the company’s share price.

Basically, they need a new revenue stream.

If you aren’t familiar with JD, they are a massive company, ranking 47th on the Fortune Global 500 list. It is one of the two major B2C online retailers in China.

In 2023, they had over $150B (US) in revenue.

They are large, and they are profitable.

So their move to get into the food delivery market by launching a zero platform commissions for 12 months—directly attacks a key industry pain point.

Traditional platforms charge restaurants 6-30% per order, eroding thin profit margins for restaurants.

Analysts suggest this model could temporarily reduce JD’s profitability but positions the company to capture high-value merchants who may agree to future revenue-sharing arrangements post the incentive period.

Backing the launch is a two tier delivery system. Restaurants can self-deliver or use JD’s Dada Now fleet of 1.3 million riders, who achieve average delivery times under 30 minutes.

Adding fast food delivery rounds out JD’s offer in the market. They already sell typical eCommerce goods and they also do grocery (see similarities to another massive online leader…). This allows them to spin their flywheel even more by creating an opportunity for even more cross selling.

I share this frame because too many of the service providers that I talk to are unwilling to take bigger steps out of what they are doing.

What they don’t seem to want to believe unfortunately is that this ‘strategy’ is severely limiting their future attractiveness in a market that has brands needed to connect with consumers in every possible way.

So the next time a new idea crosses your desk, spend a bit more time thinking about how you could make it work and what other benefits it can create for your business before saying no.

JD (or Amazon) never expected to get into the restaurant delivery business or ecosystem. But both of them understand that if consumers are there, it’s another way to stay sticky.

And in a world that’s battling for everyone’s attention, the more reasons someone can think about your brand or service, the better.

By Being Able To Get More Quotes, You Are Able To Best Manage Your Opportunities

I’m going to build out on a post I made last week about data masking.

Here’s the key message:

Your ability to not send the type of data you need to is holding you back

No matter what solution or tool or service you might be looking at, the provider / vendor is going to want data.

And honestly, the more you can give them, the better it will be for everyone.

But this usually causes a lot of brands to hit the brakes because of the fact that no one is that comfortable giving out their information to everyone.

This is a good thing to be weary about.

I can confirm that there’s often a lot more “about your business” in that data that many people think.

But it also causes you a problem, because you then aren’t necessarily getting exposed to solutions that could dramatically improve your business.

Because of this, one thing that I do for my clients and encourage everyone to have, is a masked and anonymized dataset that you can send to these vendors to get their analysis and proposals that DOESN’T give away too much about your business.

I also recommend that you create different profiles of your data.

Examples of this could be a paired down volume and mix file that only represents what would would be willing to move to a new provider or service.

It can be the data from only specific regions or countries where you are looking for new solutions.

It might even be a modeled dataset with new sales or growth built in based on your historic sales distribution (I do caution however to be VERY careful with this type of dataset - especially when negotiating rates or solutions that are volume driven).

The great thing about masked data is that you can also fragment information in a way that can further sub-segment or divide your activity in a way that doesn’t show as easily your strongest products, your key customers or your current level of diversification or solutionning.

The lookup IDs that you create are managed by you from your original data. This means that when you get solutions back from a potential new tools or vendor, you are able to quickly marry that data back up and see how it would correlate to different parts of your business.

That’s it for this week. Thanks for being here.