I Want To Replace Myself With AI
Because everyone deserves top marketing analytics expertise.
For over twenty years, I’ve had a hard time explaining what I do to anyone who isn’t responsible for spending money on marketing.
Thanks to AI, I can now explain it to anyone.
I’m a human GPT for Google Analytics; as well as for data from Salesforce, Google Ads, SEO, Instagram, Facebook, TikTok, Pinterest, Podcasts, point of sale, and web logs. I take data and help senior executives turn it into action that makes money.
I am an alchemist who transforms marketing data into gold.
Now imagine that you had an alchemist. A person who could quite literally turn lead into gold. Would you ask them to go find the lead, extract it, package it up, transport it, store it, and unbox it?
Hell no. You’d want that person turning lead into gold as fast as they could.
And yet, that’s exactly what my life -and the life of all us great marketing analysts-is like. We have to spend at least half our time working putting data into a format we can use to perform our alchemy; such as building data warehouses, piecing unstructured data together, normalizing it; and more, much more.
I’m an expensive resource, because there are not enough of me to go around. I have to turn away promising and valuable small businesses as clients because I can’t afford to work with them. It’s not fair that only big companies can benefit from my expertise. But the truth remains: lots of people can find the lead, but there are not enough alchemists in the world.
I want to replace myself with AI because a) there’s so much more gold I could create from the same lead given AI’s capabilities, and because b) I believe every company in the world-small or large- deserves to have someone with my decades of expertise and training helping their marketing team turn lead into gold.
But.
There is currently no safe or accurate way to layer ChatGPT onto Google Analytics for most companies, no matter what the startups are telling you. The core issue lies less in data security (though it’s still an issue because this sensitive information), and more in data availability. The fact is most “Marketing Analytics AI” companies don’t have enough primary data to train safe AND effective models. And most companies don’t have enough data to do it themselves, because they’re just one company.
Even gigantic companies with lots of data will struggle with this problem, because there aren’t enough AI and ML engineers available to solve this problem on a one-off basis. Given current costs, it will be hard to justify the return on investment for individual companies unless they’re spending $50M or more on advertising a year. And that assumes higher ROI/NPV projects aren’t pushing it down on the priority list. Let’s face it, marketing rarely gets R&D budget. :/
This leaves anyone attempting to solve this problem with two options that have significant challenges:
Synthetic Data Creation: While generating data based on patterns is cheaper and easier, it falls short without domain experts like me to validate real-world applicability and catch training data flaws. As I said above, we’re in short supply, and we are not cheap.
Data Acquisition from Multiple Sources: Trust is a significant barrier, as few companies can secure enough primary data from various businesses to train robust models.
As a result, I believe that at this time, only three companies currently possess the primary marketing data required to replace me at scale.
Google - due to the broad and deep information it has on many sizes of businesses, including conversion information; plus Google Ads and Performance Marketing, plus email. I’d be shocked if they weren’t thinking about it-but given the current existential threat to search, I would expect this to be de-prioritized. They don’t monetize GA that well, and it would be hard for them to see how valuable this could be.
Adobe - For similar reasons to Google, it has incredible primary data in Analytics. I worry that as a “Creative Cloud” company it too may struggle to see the opportunity in this part of its business.
Salesforce - I have incredible hopes for them. Tableau is the obvious angle here, but their massive data sets in Commerce Cloud and Marketing Cloud; and relentless focus on the right customers could prove to be a real player.
*see footnote for explanation of who I left out. (And hey, it’s the internet, so I expect you’ll school me in the comments. Bring it on.)
So What?
If you're considering investing in a tool to replace a skilled analyst like me with AI, you might end up wasting your money. The technology isn't there yet unless the company in question has access to massive, diverse data sets.
If you work for Google, Salesforce, or Adobe and are looking for domain expertise to leverage your data for AI-driven insights, let's connect. I want to help you help me replace myself with AI. This is a generational opportunity to unlock shareholder value. It is not a technically hard problem to solve.
If you're not one of these three but believe you're genuinely solving this problem, please reach out. I’m happy to be wrong, so I can focus on my alchemy.
* Why not ______?
-Microsoft - I may be proven very wrong here, but based on what I know right now; it has no significant marketing analytics assets. LinkedIn is a fraction of the story. Only fools count Microsoft out of any market, though. Perhaps with the Azure and cloud products there’s an angle, but again, customers would have to turn it over. I don’t yet see the path, yet.
-Amazon only ha ad data for its merchants and its stores
-Facebook can only use its own data, same problem that LinkedIn has. EDIT: 8/23: With Meta’s announcement about Google Analytics Integration, they might be able to change my mind….
-ChatGPT can only do this if they’re totally violating their user agreements, and given their broad ambitions, I don’t see this as making the cut in terms of importance. (see also Anthropic but it can’t do data so 🤷🏻♀️)
-Did I miss anyone?
--EDIT: Yes, of course, I missed Apple. Like Microsoft, only fools count them out of any game. They have enough cash to buy their way in. But in this case, as things stand today, they too have no marketing analytics data to speak of. They have their own, of course, and it’s not nothing. But to build this tool at scale, they’d need to find a way to get more of it.
* AI Disclosure: I used ChatGPT4o to help me with reviewing and editing this piece. I came up with the alchemist analogy.