Adjusting The Backward Thinking Around AI
Marketers have a missed opportunity around what John Connors, founder and president of Boathouse, calls performance artificial intelligence (AI). It's a focus that brands and agencies are missing out on. Generative gets all the attention, the shiny object, but "it's just a fancy bar trick."
Performance relates to strategy and using AI to drive actual results and business outcomes. Most in marketing and advertising are focused on using generative AI to write social copy or a press release faster, which is helpful, but that's barely touching the surface of AIs potential.
Performance AI needs to be the future that marketers obsess about, according to Connors, who grew up in the advertising industry, starting his career at Hill Holliday.
Connors was formerly CEO of Zentropy Partners, an internet services business, and as a part of the McCann World Group Management team. Then he founded Boathouse in 2001.
Through an exchange of emails, Connors shared his thoughts with Inside Performance on that missed opportunity. Here are some excepts from the exchange.
Inside Performance: What are brands getting wrong when they think/use AI?
John Connors: As the AI arms race begins, too many brands are using AI tactically vs. strategically. They are focused on TV ads, press releases, banner ads and emails. These are not things the CEO or the board really care about. These tactics will rarely impact the business significantly. The opportunity is to look at how AI can transform the entire marketing department to drive more speed, more growth, more innovation. If we can translate AI to business impact we can get the CEO, board and market to care.
IP: How should they be thinking about using AI?
Connors: First, we need to get beyond tactical creative A.I. (a.k.a. “shiny objects). There is strategy AI, media AI, targeting AI, analytics AI…all in addition to creative AI. Creativity is approximately 5% of a marketer’s overall budget but it is getting 95% of the attention.
Second, AI should force a fundamental re-evaluation of how we have set up marketing departments and agencies. Over the last 25 years we have all become a collection of siloed specialists in response to digital/social/mobile channel proliferation. AI does not respect silos so marketers and agencies will be forced to create models that break down the silos.
IP: What type of AI are brands using most and why?
Connors: The only AI requests I receive from clients are for creative. Brands and agencies are obsessed with creative AI as evidenced by all the silly news coming out of Cannes. I think it is interesting to note that while agencies are hyping A.I., the stock prices of the major holding companies are in decline. To me, this is a signal that agencies and brands have not cracked the growth code on AI.
Equally fascinating to me is that the “gold rush fever/bubble” is real as everyone races to launch their proprietary creative platform. This seems absurd to me…bad business. The barriers to entry are real. Microsoft has invested $10 billion in ChatGPT, and VCs are pouring money into the space. It is a great time to try all the tech and lean into the winners…not the time to be deluded and think we are tech companies/engineers.
IP: What is the cost for a company to implement AI?
Connors: Like everything, you can spend a lot, or you can spend a little. You can spend big on enterprise-type A.I. with category leaders in strategy, media, creative, analytics -- or you can try, test, experiment, learn with the AI challenger brands.
But making a big, enterprise-level decision with no experience does not seem smart to me.
Like a 16-year-old driver buying a Ferrari. I would recommend everyone make multiple $50,000 to 75,000 bets to test the different technologies and to test their organization’s ability to work across silos. From those tests, brands and agencies can then move to build an A.I. stack that is purpose built to deliver business results.
IP: How to ensure copyright protection on content, so large language models do not infringe?
Connors: The copyright issues are a moving target, as we are seeing in the media.
The first step is to leverage tools that can provide assurance that their data sets are “clean” and free of copyright issues, bias, etc.
The second step is to have a good lawyer. The best firms are moving aggressively into the space, and we are very deliberate about making sure we have vetted the tools with our legal advisors before we use them for clients.
The risks are simply too big for us to be cavalier with our client brands.