An Emerging but Obscure AI Play in London
And the hidden threat it could pose to hundreds of stocks
Lunch was sushi, sashimi, chicken caesar salad, bruschetta, quiche, broccoli and bean salad, squash salad and more – a confusing cocktail, and a portent for the afternoon.
I was on the top floor of WPP’s impressive waterfront offices, 10 days ago, for their Capital Markets Day. There were over 100 people in the room for 4½ hours of presentations from 13 WPP managers (10 men, 3 women) and 1 customer.
There were just two 10-minute breaks, with the second curtailed to a 5 minute stretch. It was a long afternoon.
Smooth talking 50-something CEO, Mark Read, kicked off with a polished presentation (auto-prompters were used throughout). He set out the WPP’s recent history and plan under the heading “Innovating to Lead.”
It took him under 15 seconds to mention AI and my heart sank.
But as we went through the afternoon, it became apparent that WPP has made real advances in the application of AI.
We’ll get onto that in a second, but first a note for the professional analysts and fund managers among you: my next Forensic Analysis Bootcamp starts on March 4. One world-famous institution has sent six (!) of their analysts on the course. For more details and a full schedule of workshop topics, please go here.
WPP’s Plan
Read talked of three periods for WPP:
2018-2020 Stabilising the business
2020-2023 Accelerating growth
2024 on Innovating to lead
Now if you go to an advertising agency’s investor day, you should expect to be sold to and it’s bound to be good. So I went into this meeting with even more than my usual dose of cynicism. Yet I was still impressed by this one statistic - a number that, in my opinion, should make customers queue around the block for WPP’s offering. WPP claimed:
they can guarantee clients a 25% increase in conversion on their website product detail pages
Because they are actually achieving a 40-50% uplift.
I thought that was an unbelievable statistic and one person I mentioned it to thought it was a lie. I was astonished that nobody asked about it – not a single question. Of course the analysts were all focused on the guidance - the range of operating margins and revenue growth predicted. Almost all the questions tried to get behind those data points instead of asking about this, which surely must be the real question.
For those less familiar with marketing buzzwords, a product detail page looks like this:
Product Detail Page
Source: Amazon
WPP used the example of mascara and clearly it’s going to be most readily applied to FMCG type products but a guarantee of this nature seems astonishing to me – I would kill for a 25% improvement in conversion rates on my website and would happily share the upside with the agency.
This raises all sorts of associated questions but the one I asked the CTO, Stephan Pretorius, in private was why there wasn’t a queue of customers outside. He said it’s very early days and obviously this type of initiative would have a long lead time. I then wonder why announce it to the market at an early stage as you are also warning your competition, but I guess this is an industry in which things don’t stay quiet for long.
Other important questions include:
Where are the competitors? Publicis did an AI presentation the previous week and that talked of an astonishing 45,000 engineers.
Assuming WPP has an advantage, how long can it retain this?
A 25% uplift is amazing (let alone 40-50%) but it must be a one-off. Future benefits will be much lower, even if WPP can retain any advantage.
This is a zero-sum game. Competitors to WPP’s clients would respond to loss of market share.
Competitors can train their LLMs on WPP’s successful product pages and that might be one way of catching up.
I am sure there are many more and I am no expert in this field but it raises an important broader question. Recent improvements in AI are clearly a huge development but we are in the early stages of its implementation. If AI is as big as, or bigger than, the internet and mobile, should we not expect it to be every bit as disruptive? And will it not be even harder to predict?
If WPP is successful in this initiative, its clients will see an acceleration in sales growth and their competitors will slow – imagine the relative impact on two companies’ share prices. Perhaps investors are totally clued up on all this, but I doubt it. More likely, this could introduce a random and even less predictable element into companies’ performance – should that not make us more wary of paying up too much for past performance and perceived quality.
This was my most important takeaway from the afternoon. While Nvidia has been the AI darling, I have seen much less discussion of the second and third tier impacts of AI, beyond issues like the impact on banks’ profitability by stripping out cost.
Paying subscribers can read on for more on WPP’s AI plans, other business segments and an initial look at their stock (over 1000 words in all).