Product Positioning, AI For Good, and Google’s AI Strategy
An idea about *how* to create the next wave of AI products, some actual positive-impact product ideas, and a hope about how Google will enable the AI ecosystem
Here we are, at the beginning of a big wave: AI-powered services are coming fast. ChatGPT and MidJourney are just the opening salvo, the very beginning.
This is an exciting moment to be an idealistic entrepreneur.
I believe these new AI superpowers can help people live healthier, safer, more prosperous and more fulfilling lives. I believe some of these tools can help us live sustainably and comfortably on this planet.
But we’re not there yet. The products and services that use the new advances in AI are just getting started, and at this early stage it’s not totally clear which products will actually take off, which ones will be novelties that fade quickly, which ones will never get any market adoption. The race is just starting now.
I have some thoughts to share about the AI products that I’d love to see, and how I hope Google in particular will enable the AI ecosystem. But my thinking starts with the methodology we use to create these products in the first place.
* * *
One of my favorite books about innovation is actually about marketing strategy. Written in 1980, Positioning: The Battle For Your Mind is a classic, still taught in marketing classes around the world today.
The authors were advertising executives from the 1950’s and 60’s, the era of Mad Men. They were faced with a problem: everyone was spending more time with TV and radio, so people were increasingly bombarded with media and advertising. Everyone started to develop a reflex to tune out new advertisements, to ignore them, since all these new advertisements sounded so similar, anyway. (This, obviously, is 10x more true today!)
The challenge was this: How to create a marketing campaign that would stand out, that would cause people to stop, notice, and think, “huh, that does sound interesting!”
Their strategy was to first understand the concepts that already exist in the minds of your potential customers, before they’ve even heard of your product or brand. Then, using those concepts, you can use one of several techniques to redefine a category in which your product is the best.
They describe the basic premise as this: “The solution to a positioning problem is usually found in the prospect’s mind, not in the product.” This is a fundamentally humble approach that requires real research and listening. It is about identifying a problem that a customer needs to solve, but also about understanding all of the concepts that a customer already has in their head about the context of the problem, and what possible solutions exist, and the set of expectations they have about how to use those other solutions.
Then the basic approach to finding a strategy is this: “A positioning exercise is a search for the obvious. Those are the easiest concepts to communicate because they make the most sense to the recipient of a message. Unfortunately, obvious concepts are also the most difficult to recognize and to sell.”
I love this approach.
Normally, Positioning is something you do after there is a product or service to sell. Some of the classic case studies are about how to market a tourist destination, or a new car.
But at Imagination Machine, we turn the traditional process upside-down. We start with a positioning exercise, before there is any product or service that exists. We don’t use it as a marketing exercise, because there is nothing yet to market; we use it as a product discovery technique.
We start with a problem we want to solve, or a hypothesis of some sort. Then the first step is to find the “obvious” solution to the positioning exercise, using concepts that already exist in peoples’ heads. If we properly understand the problem, the context, and the existing solutions that are already in the popular consciousness, then we can find the right niche or new category for a solution.
Then, as a second step, we use that positioning-solution to build a product or service that fits the description. And in building that product or service, we frequently need to develop technologies that do not yet exist.
The classic innovations approach is this:
new scientific or technological innovation > (2) product development > (3) marketing positioning.
Our approach is often this:
positioning > (2) product development > (3) technological innovation.
When we co-created Beem, with Ralph, Pierre-Emmanuel and Arthur, we used this approach to pioneer the new category of “plug & play solar”. And not only did it work – market adoption has been amazing – but it also led to real tech innovation to enable the product development, with a world-class engineering team and patent portfolio behind the scenes.
Of course each project is different, and we use various methodologies depending on the circumstances. Some of the more modern “Product discovery” methodology tools we love to use at IM, like Jobs To Be Done and Discovery Discipline, can be seen as more specific frameworks that build on the underlying ideas of Positioning.
These methodology ideas are in my mind when I think about the latest advances in AI.
* * *
With Positioning, you have to answer some basic questions, like:
What kind of thing is this product, and how do I use it?
What is it for, and when should I use it?
Who is it for?
Why is it special?
The latest generation of AI tools have come to market, but it seems to me like they are technologies, but not yet products. ChatGPT and MidJourney are super cool, and extremely useful, but they don’t propose answers to most of the key positioning questions.
I believe that without answering these questions – and building the corresponding product and tech innovations – the latest technologies will not be adopted by the masses.
Last week, the NYTimes chronicled 35 Ways Real People Are Using A.I. Right Now. The most striking thing about the article was that it needed to be written at all. For most people, I suspect, it’s not obvious what these tools really are, how to use them, or why to use them.
If I’m correct, then ChatGPT will not be the next Google, unless they move quickly to define key use cases, or personas, and develop custom features for each of them (which they might).
It is more likely, I think, that a more specific class of products will emerge, each one tailored to a certain problem and use-case – that will all be built on top of technologies like ChatGPT or its competitors – and it’s these more specific apps and services that will gain widespread adoption.
As a trivial example, consider writing custom books, which is one of the 35 use-cases from the NYT article. Someone used ChatGPT to write a “a custom book of cocktails based on the tenets of traditional Chinese medicine written in the style of the J. Peterman catalog”, with illustrations, as a gift for Valentine’s Day. If you start with Positioning, and design a product just for that kind of usage, then you might end up with an app with a specific name that suggests creating beautiful, one-of-a-kind books; you would create an interface with a bit more hand-holding so that someone could choose the style of writing, go through some exercises to describe the kind of content desired, etc. And perhaps the technology behind the interface is just a large language model like ChatGPT and an image-generator like MidJourney; or perhaps, to deliver a truly superior experience and meet the demands of the positioning solution, you need to build a custom data-set to augment the generic tools. You probably need to build some technology to automatically layout the book. You get the idea.
I’m interested in using these AI technologies to make a positive impact in the world. I can imagine specific apps, each with their own branding, interface, and technology, that address some of these opportunities:
Helping farmers with the daily troubleshooting that comes up – pests, plant illnesses, unexpected weather, machinery problems – and suggesting, by default, organic and nature-based solutions. This same AI could help farmers plan cover crop rotations, design hedges, and implement other techniques of regenerative agriculture that increase yield while improving the planet.
Helping teachers with grading homework and filling out administrative work, and suggesting lesson plans that are personalized to the interests and learning styles of each student.
Helping healthcare workers with the heavy administrative work and note-taking necessary in their daily work, so they have more time to spend with people who need care.
People are already using ChatGPT to do these things, but that requires too much effort for widespread adoption. Perhaps with a layer of custom UX in front, and a layer of customer domain-knowledge in back, you can build a world-class product.
And of course there are many, many other examples, these are just three that we’ve been thinking about a lot at Imagination Machine. (If you’re interested in working on any of these things, please get in touch!)
I bet that if you start with a deep Positioning exercise, you would find great product concepts, discover the custom UI needed for each, and that would drive some technology advances. Perhaps each of these concepts would require new, custom data-sets to plug into the generic AI tools. Perhaps the new apps or startups could find ways to share revenue with the people who contribute to the custom data-set, as Jaron Lanier suggests in his article in the New Yorker last week.
* * *
I have been thinking about the battle between OpenAI, the company behind ChatGPT, and Google, which just came out with a competitor to ChatGPT called Bard.
ChatGPT has a great product name that solves some (not all) of the key positioning problems.
How do you use it? You chat with it.
What is it? Well, I would bet for most people, “GPT” sounds like a robot name, like C-3PO from Star Wars. It’s an intelligent robot.
I would guess that in mainstream market tests, people would naturally understand that ChatGPT is an intelligent robot you chat with. That’s already a big step.
Google Bard, by contrast, feels like a weaker name to me because it doesn’t do any of the positioning work that is probably necessary. “The Bard” is a nickname for Shakespeare, so I can see why it was chosen as a brand for a chatbot that aspires to be eloquent and prolific. But I would bet that only a very small minority of people are familiar with that definition of Bard, or that describing the technology is the right move in any case. Better to start with what is already in the heads of the potential users.
If I was still working at Google, I would advocate for a product strategy that uses the AI tools, like large language models, to power a vast set of apps. Google itself could develop some of the key apps (the same way that they build some of the key apps for Android, like Maps, Photos, Mail, etc., while also enabling an ecosystem of millions of others). Product positioning would be a great way to design those apps.
Imagine ten great Google AI-powered apps, each one with a suggestive brand that lets it lead a new category:
an academic research assistant that uses the data from Google Scholar
a personal advice app, or a personal health assistant
an assistant for software developers, like a competitor to Github Copilot that you can also chat with
a travel agent that uses data from Google Maps and reviews
etc. etc.
Each one of these apps would have some custom features and clear use cases. Each one could also have power-features that are available if you subscribe to a premium plan, a natural revenue model for these use-case-specific AI tools. (And all the premium plans could be bundled via Google One.)
Google could also build a great no-code AI App Creator, so that millions of people could create their own apps using the same technologies.
Then, instead of building a new destination like Bard, those apps – both Google’s flagship ones and the platform of user-created apps – could start showing up directly in Google Search.
If that ecosystem existed, it would help me as a positive-impact entrepreneur to have the tools to develop the apps I have in mind, and a way to distribute them to new users (via Google search).
Perhaps some of these ideas, or better ones, are already in the works at Google.
* * *
In any case, we can be sure that a massive wave of product innovation is coming.
I hope that it will make the world a better place.
Most of all, I hope that idealistic entrepreneurs seize the moment. This is our chance to shape the future.
Thanks Rob! The positioning framework is definitely valuable, avoiding the tendency to be blinded by the brilliance of new technology.
I perhaps disagree on one point - while you're right that google and others will create value by integrating AI in a seamless way into other tools, I think there is great value in having an all-in-one go-to AI, and right now OpenAI is winning at that. So they might become "the next Google" of that space, even as they don't replace google for other tasks.
Bard is Google's inevitable attempt to prevent that. It seems unlikely to work for the moment, unless they introduce some major breakthrough soon, and I'm not sure integrating with other google products is enough of a breakthrough. Google no-code AI app creator is a great idea, something like that might be big, but at some point creating a no-code app starts to blend with the ability of ChatGPT+integrations to do whatever you want.
Thanks Rob for a great insight on what AI could possibly lead to in the future. As a marketing lover and sometimes professional, educated both in France and the US, I really appreciated IM's positionning approach. It should always be this way around if you want to really provide the customer with a product or service that really fills a blank, and therefore is useful. Would love to talk to you one of these days. Kind regards, Eric