'AI is trivial in our business... the bigger picture makes me nervous': Marco Rimini

A month after Sir Martin Sorrell left WPP we caught up with Mindshare’s global chief development officer, Marco Rimini.

 

Rimini described that as a difficult time and what added to his woes was the exit of Arsene Wenger from the football club he supports, Arsenal. “Imagine how I felt! Two father figures left me in the same week. It was quite difficult for me to take in.”

 

Rimini had joined Mindshare in 2006 as head of strategy. We caught up with him while he was in India last month to learn more about how the industry has changed in the last 12 years, the emergence of AI and what does he make of the Indian offerings of Mindshare.

 

Edited excerpts:

 

How has your role changed since 2006?

 

I joined originally as head of strategy in 2006. I then became head of the worldwide group which runs all of the global clients. It was a CEO and P&L role. It was actually more of a product development in a P&L role. So in that role, I not only managed the client teams, but all the services to support the global clients.

 

I created ‘FAST’ out of London, which is social, programmatic and SEO put in one place and grew that from two or three people to more than a hundred people. I also run the data and tech team, business planning team for a while.

 

So I went on a journey from strategy to CEO in digital products to now a role in overall product development and overall corporate development. I can do that and I feel comfortable because I did the CEO role earlier. 

 

The reason that it makes a difference is that if you’re doing the CEO role, one is understanding the commercials of the business and the clients talk to you differently. If you’re the strategy guy, they talk about the interesting stuff, the stuff they think you’re interested in; they don’t bore you. If you’re the CEO, they talk about everything, so you get to know the clients more intimately and practically. And so I’m trying to get the commercial and strategy piece together now. 

 

It’s an interesting time now because so much is changing. I’ve changed my role, and that’s one of the good things about Mindshare. The company gives you the freedom to change what you do.

It’s now global development officer. In this context, I’m looking at how we develop our product, services and the way we do corporate development. 

 

We have a separate new business development team.  I’m trying to look at the new stuff.

 

How much of the ‘new stuff’ is digital led?

 

Well, yes a lot of it is definitely digitally led. But, more of it is performance led. I’m a big believer that performance is the single most important issue today.

 

And then I focus on data and technology. That is the infrastructure that will then support all of those other ‘digital led’ things. You sort of need to put the performance team and the performance mentality in the centre of agencies. 

 

If you listen to the global clients in particular, and one of the reasons I’m here in India right now is to listen to clients here, they all worry about Amazon and e-commerce. Strategically, that’s a bigger worry for them right now than what their brand stands for. They can be out of business because of these new emerging businesses and then there’s no point worrying about what your brand stands for or redefining it! 

 

They’re all interested in how content fits into the performance and e-com world. They are also interested in the global/local way of organising that. In the platform world you can do a lot more in the regional level than at a global level. 

 

How is India performing as a market for Mindshare? What would be the agency’s strengths in his market?

 

I have learnt and listened a lot to the global clients because that’s what my job was. I can tell you a lot about them. I haven’t spent so much time in the market though. I’m visiting China (have already done before Christmas), India and USA as these are the markets I wanted to get closest to partly because I’m inspired by these markets. 

 

Mindshare India has always been a company that has gone first in a lot of development things. It’s always at great scale and that has helped. Content has always been a strong play here. From gaming to the development of YouTube content, to development of long-term content, this office has been way ahead of anywhere else in our (Mindshare) world. The closest to India in terms of content would be China, where in particular in gaming, the Chinese are very good because of all their platforms. 

 

India is an out and out leader for Mindshare in terms of content. I’m very interested in learning this.  The strategy we have here is ‘have these interesting relationships’ with outfits. 

 

With The Glitch, we have a content partnership and with their new product called ‘The flux’ they are getting into data plus content. We are open source and we work not just with WPP companies but outside as well with companies like FoxyMoron. We have a close relation with them. We work with other suppliers too. I’m trying to drive an idea, that black and white in the new agency model, where you’re either owned by the agency or you’re not owned by the agency, is not very important thing to clients anymore. We work closely with Adobe too. In terms of the future of the agency – if you just view it as the people you own, you’re going to limit the size of the agency and limit what we can do for clients. India is particularly good with the tech and digital partnerships. 

 

While India is strong in content, in which area do you think it’s not that strong?

 

I think it’s only starting now - the journey on data and tech consulting. That’s partly because the clients are only looking at it now. You have got only three per cent of your business in e-commerce, whereas some markets are at 20 per cent. So, the management of e-commerce and the problems related to that are only just starting right now. So, that’s not as mature as some markets like the United States.

 

With mobile penetration increasing, the rural population will soon be on apps and e-commerce. I’ve seen how stuff like Instagram can work with regards to e-commerce and that’s really interesting. 

 

You’ve written a lot on AI… What’s your view on it? Will it help humans or destroy? 

 

Humans create AI, so it should really help them. There’s a great book called ‘Mutually Assured Destruction’ or MAD. It gives great examples of how algorithms have led to terrible decisions. She has then become an advocate to really control these algorithms and AI. 

 

In fact there are quite a lot of people that are saying AI should be opened up to public scrutiny. Especially the ones that are being used in the public sector. 

 

So, to answer your question, I suspect we will have to have a structure where some bodies will have to be involved to make sure that AI services humans. 

 

The UK is a good example where there was a contract between a hospital group in the UK and Google Deep Mind AI group. I’m sure it was fine and they gave hospital data to Deep Mind and that was in turn going to help solve diseases seeing patterns in data. I’m sure, all sides of that was the perfectly good intention, but you still effectively had the government (they own health services in the UK) giving public information to a private company. That makes me nervous. 

 

The other example given is that AI is supplied to criminal records and the identification of positive profiling. What it leads to is an over index in the States of people of the scrutiny of a certain race. And that’s not great. 

 

In our business, ‘AI is trivial’. 

 

So, keeping it to the trivial advertising and media businesses, will it be taking jobs away?

 

It will re-cast jobs for sure. Programmatic is a classic example. We have a product called Anna. It helps us get 70,000 variables rather than 40,000 variables that a human can do through a programmatic campaign. But, what it means effectively is that we can have the same person doing more. 

 

What will then happen is that you can do greater sophistication. We can do much more. The moment a client pays us a certain amount of money to do a programmatic campaign, we can only do a certain amount of optimisations. Say, they pay us a 100 pounds for those optimisations. In the future we’ll be able to do 500 optimisations for the same amount of money. So, I have no fear it’ll take away too many jobs. 

 

If we get outcome based remunerations, then we’re in a different ball game altogether; they take some risks, we take some risks. That’s why performance in e-commerce is so useful, because with that leads directly to an outcome. So if you look at the acquisition and performance clients, they’re quite happy to pay by outcomes. It’s the brand clients that get all worried about it. 

 

For acquisition clients, if we lower their cost per acquisition, they’re happy to give the share. So, we work with Booking.com. They’re very happy to get involved in these conversations because if they get 10 pounds and we give them 15 instead, they don’t really mind. So, that’s a good journey to take commercially, and AI will help us with this. We’ll probably partner a lot of people to do this on the back end work.

 

I’m interested in AI, and if you look at the company and you’ll find us sitting next to someone today who has a masters in communication, business or media, and suddenly I’m thinking are we making the most of this person’s talent? 

 

If we can get AI/tech involved more, we’ll have fewer people playing around with Excel and that’ll make the business more interesting. 

 

AI can easily destroy journalism too with the same logic. So, I see where that question came from, but hopefully you won’t let it. I used to work with The Economist a thousand years ago. I was working in the information sector during my university holidays. My job was to help get information for the journalists. This was pre-internet. We used to cut the newspapers and file the cuttings. Journalists would come and ask us for information and we would go get the cuttings. I was a reasonably intelligent person, but that’s what I was doing. Now, when you research a story, you’re using technology and AI to make the process quicker.

 

You can therefore argue journalism has got worse, because all one is doing is Googling other people’s writing and then slyly changing it. Or you can say it’s got better because the quality of the research has improved. In five minutes one can find out what used to take us weeks earlier. It’s not AI that has destroyed journalism; it’s the advertising revenue model that is screwing it. AI has helped journalism and it should help us too.

 

How has AI impacted other markets? You were in China recently. What’s the scene there?   

 

The tech scene is completely different because of the domination of the platforms like TenCent, Alibaba etc. To manage the market and what they do is quite specific.

 

The amazing thing with China is how fast they’re going to payment by device. They have gone so fast in that direction. If you take out a card in China, you feel like an idiot. No one pays by card. Cash is still okay there. Otherwise you have to pay using your phone. 

 

I don’t pretend to know about the other China as much as the commercial one and they’re growing really fast.