E-Commerce

Wattle Labs has created a "Large Behaviour Model" (LBM) of the human population. The LBM, an analogue of LLM's, which seeks to understand human behaviour, rather than language.

This video explains how our technology can can enhance E-Commerce.

#datascience #AI #digitalmarketing #ecommerce #behaviourchange #BehaviouralScience #psychometrics #nudging #boosts

https://www.youtube.com/watch?v=AJ3tymZgQss

Transcript

Hello e-commerce people, my name is Aaron and this is Wattle Labs. The Wattle Labs large behaviour model can give companies an advantage in three ways. Our model can be used to enhance the UX with dynamic layouts, menus, and lists based on the specific customer who is on the store at that moment. We can then present dynamic offers based on the customer's persona. And finally, we can personalise product imagery and descriptions to target each and every customer's latent psychological profile.

In the first example, there's a carousel of eight offers. I'm going to pretend that I'm a very high income person — even though normal me loves Dove, high income me goes to some Bath World with bath bombs and essential oils from South America. By presenting an offer at the front that isn't psychologically — or in this case socio-economically — aligned with me, one eighth of the revenue that this user interface element should be generating is lost. Not only is that revenue lost, there's also the opportunity cost of not getting the right offer in front of the person. There's zero revenue coming from the wrong offer, but it's also minus one eighth of revenue because of the sale you didn't make. That's two eighths — one quarter — which is a 25% loss of revenue based on this user interface alone.

Instead of having a bunch of static offers, what we would suggest is that you create a huge number of offers — far more than eight, perhaps a hundred or more. Just keep adding offers. We can then psychologically profile each offer to create a persona, and we'll have a list of personas so that when high income Aaron comes onto the screen, we can assemble the eight offers that are most psychologically aligned with me personally. This means you're using screen real estate and customer attention to the maximum possible advantage, which then has the result of maximum dollars.

The next example is a simple product offer. In this case the offer is half price. What we'd suggest is that offers shouldn't just be statically half price — we can psychologically profile different offer types, like 50% off, and link the best offer type to each person individually. The second issue with this example is that it has six products in the carousel. I'm going to be healthy me here — so I'm not going to be purchasing chocolate. Around 60% of that carousel wasn't aligned with me for various reasons, so that's 60% lost revenue, plus a further 60% in opportunity cost from not putting the right offer in front of healthy me at the right time.

So what we can do is take all of your sales data — all 10,000 products — profile it to create personas, and then match those personas to healthy me. The result is that every person sees a unique list of six offers on screen that most resonate with them on a psychological and behavioural level, which should maximise revenue for every piece of screen real estate.

The next example has the same issues — we can psychologically profile the different offer types and make sure that the type of offer presented to each person is the one they're most aligned with. Healthy me loves three of the things shown, but definitely doesn't want anything air fried. So again, we're down minus 40% revenue for that user interface element — 40% of revenue you're not going to get because people won't click on it — plus minus 40% for the things you should have offered instead.

In the following example it's essentially the same concept but applied to product groups — fruit and veg, deli meats, and pantry. What we can do here is combine all the fruit and veg sales and profile them in total to create a persona for that group, then match that persona to healthy me. And we can do that across all 10,000 products and all 10 million customers.

In this next example, the content involves clickable buttons for things like recipes. The same construct applies — we can take all of your content, profile your web traffic, create a persona for each page, and then present the right content recommendation to each individual user. This isn't necessarily an opportunity to make money directly, so there's no lost opportunity cost in the same sense, but it's definitely an opportunity to increase customer engagement, satisfaction, and understanding of the products and services the retailer offers.

The next example, which I love, is a Polaroid camera. We can look at the psychological drivers of the people buying it, and there might be quite different drivers across groups. There are the oldies — like me — who want to relive their youth and childhood, and then there are the kids, anyone under 40, who want to be retro cool. In this example, we can take the product image and effectively adjust it to target the psychology of each group. We can also modify the product description to align with one group or the other. Our system doesn't actually use pre-baked groups or personas like that — we have a virtually infinite number of personas, because everyone is unique. The net effect is that the persona is tailored conceptually to every single person, so every individual visiting the store sees a product description tailored to their specific circumstances.

In summary, what we're really advocating is this: instead of having a one-size-fits-all online shop and hoping that people interact with it — spending a lot of time trying to figure out what people do and why — you can use psychological profiling techniques to take your one conceptual store and allow our behavioural models to slightly adjust it, optimising it to create a unique experience for each person who comes onto the site. You're providing unique shopping experiences for every customer. It's basically taking your static store as you've built it, and supercharging it so that you're effectively creating a unique experience — in some respect, a unique store — for each and every person.

Anyway, that's how Wattle Labs could be used to enhance e-commerce. That's all I want to say. Thank you very much for listening. Bye.