FMCG

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 enhance FMCG sales and marketing

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

https://www.youtube.com/watch?v=ghqQ6UP9mGY

Transcript

Hello Australian FMCG companies, my name is Aaron and this is Wattle Labs. Wattle Labs has created a large behaviour model of the Australian population, covering all 28 million Australians. Our model has 60 attributes, we make 1 billion predictions, and we have been independently verified to be 92% accurate.

As a consequence of our behaviour model, we can optimise your sales network. We can look at how people and their locations interact with your stores, optimise your catchment areas, and even optimise store layouts, merchandising, and the position of different items within your store. We can also do dynamic campaigns. Instead of sending a campaign out to a million people straight away, we can send it to a small panel of, say, a thousand people first, analyse the responses, psychologically profile the behavioural drivers that caused people to interact with that campaign, and then send it out to a million people we already know are predisposed to engaging with it — before we've even hit send. We can do the same type of behavioural profiling on your apps and e-commerce stores, so that they're psychologically suited to the person looking at them, rather than presenting the same website to everyone.

We can also provide a new revenue model for Australian retailers. Currently, Mastercard provides a data product covering consumer sentiment and behaviours of the Australian population. FMCG companies, on the other hand, can provide really accurate line item data. Using Wattle Labs technology, Australian retailers can provide better insights than what they're currently purchasing elsewhere — with far greater granularity and resolution — giving much better intelligence on what the Australian population actually wants.

Finally, we can enable co-marketing and brand partnerships without sharing data. This probably comes up all the time — Telstra wants to sell phones to Everyday Rewards consumers, for example. The way this was typically done is by pooling all the data together and hoping that a bunch of data scientists can find some propensity model to figure out which Everyday Rewards customer should be shown a Galaxy ad or an iPhone ad. By using behavioural and psychological profiling instead, we can enable Telstra to sell phones to Everyday Rewards customers without exposing any sales data or PII data. We can provide a much better recommendation model, deliver obviously higher sales, and never expose shareholders, customers, or corporations to risk. The insights are far superior to old prediction models, which basically use 1990s technology. This is machine learning. This is cutting-edge sales and marketing.

So finally, I just want to pose a question to everyone in FMCG — and indeed all of retail in Australia. Everyone knows these three offers: offer one is two for one, offer two is 50% off, and offer three is buy one get one free. The effect of all three offers is the same — people get something at half price. The question I have for every single retailer in Australia is: why do people prefer each offer type? What are the psychological and behavioural drivers causing people to prefer a two for one versus a 50% off? And question two: if I picked a random customer from, say, the Everyday Rewards program, can you predict which offer type would be most effective for them?

If you can't do that, you're necessarily compelled to look at ways to help Australian families take advantage of these opportunities in the best, most seamless and frictionless way possible. I'd say you're really compelled to ask us to look at your customers and do the analysis. Let's figure out why people click buttons. Let's figure out what's really causing people to choose John West tuna over Green Seas. Let's do the work — because by doing all of that, the end result is a better customer experience, more sales, and less cost.

That's all — I'll do a quick demo. The question is: where should you put a store? We've colour coded it so green means go and red means no. There's more opportunity in Melbourne because it's a bigger green area — and I'm not a Melbourne person, so forgive me. But if you are Woolworths, your greatest market opportunities are Craigieburn, St Albans, Hoppers Crossing, and Werribee. Focusing in on these areas, if Woolworths wants to maximise revenue and EBIT profit, you should be really focused on putting a store in Manor Lakes — either right in Manor Lakes, further out, or possibly somewhere in between.

If you were to put a store there, we can predict your catchment area, we can predict the full psychological and behavioural model of the people shopping at that store, and therefore which products you should stock. We can also predict that this new store would have a catchment area of 7,684 people — that's 100% new traffic — generating around $46 million in gross revenue with an EBIT profit of $2.1 million. It would also cannibalise foot traffic at your Woolworths Wyndham Vale store by around 15%, and at your Werribee Central store by around 2.25%. But overall, when you sum up all the pluses and minuses, we predict a net increase in revenue of $21 million and a net EBIT profit of $982,000 from putting a store in that location. That's network optimisation in action.

Anyway, that's all I want to say. Hope you have a really good day, and thank you very much. Goodbye.