QuantumBlack began its operations as a data analytics company focusing on the domain of Formula 1 racing — from car design and testing, to measuring precise locations of cars on the race track within centimetres, and pit stop duration to milliseconds.
It has now extended its services to the enterprise market, serving companies that want to improve their operational performance — in diverse verticals — using data. It was purchased in 2015 by consulting giant McKinsey, and has grown rapidly since then, building a team of 300 in six offices around the globe.
In early July, it opened an office in Singapore, supported by Singapore’s Economic Development Board (EDB), to serve customers in the region. We sat down with CEO Jeremy Palmer during the launch event in a broad-ranging interview on data science, company culture, and the company’s unusual trajectory from being an F1 specialist to a broader enterprise technology company.
How has the transition been, from being independent to being a part of McKinsey?
Our space kind of looks trendy in some ways but it’s actually quite a tough space to survive in. The commercial models are not mature yet, and in the early days we were never that secure. So [now] it’s nice to be able to to look everybody in the eye and say ‘payroll is definitely going to happen this month.’ And the logic for us for combining with McKinsey is: we saw a world where we needed to have a lot more domain knowledge to be relevant, rather than being a group of pure researchers — you really need a lot of domain knowledge to be relevant and to drive impact.
We needed better access and we were hiring consultants and project managers to lead our teams and to help us get out there in front of clients and CEOs. We learned the hard way that there’s a lot of theory in this space, but the practice of driving data and analytics to real, lasting change involves a lot of organisational understanding and change management.
Change is change. It’s always a bit bumpy. But the overall logic of the situation has really helped drive us forward, we’re very fortunate to be where we are.
[Becoming part of] McKinsey was pretty obvious when you thought about it. Fortunately, I think McKinsey itself is on a journey of change anyway. The timing was good, in the sense that data and machine learning were becoming more relevant and moving faster to the centerstage of everyone’s agenda. It’s not as if we’re the only technical people at McKinsey — they’ve been hiring a lot of technical people around the world for a while anyway, and we were able to work with our new colleagues to help the broader McKinsey agenda as well.
You mentioned earlier about the commercial models not being entirely certain yet. Do you mean specifically in terms of AI, machine learning and data science companies?
Yes that is exactly what I mean. People talk about this stuff as though it has already landed, but the truth is that it has landed in relatively narrow ways.
We’ve seen people scale quickly if they have applications which lend themselves to a relatively narrow part of machine learning. We’ve started to move beyond advertising optimisation — which is what it was about for years, really — to a world where people can see opportunities to apply this stuff in almost every field. But the number of organisations that are successfully dong this stuff, broadly, on an enterprise scale (rather than deploying a narrow solution against one part of it) is very small.
How do you take such a specific case [Formula 1] with a very concrete goal in mind and apply it broadly?
The concrete goal is a really important point, but I kind of look at it the other way round and say ‘what does every project need to have? What is analytics for?’ Where a lot of people go wrong is that they’re not clear about the goal and they’re not clear about the right role for technology to play or how to fit it into their organisation.
We learned from sport that the most important thing is to be focused on measurable performance improvement. That’s highly extractable, generalisable to almost any field. Our purpose is to help organisations be the best they can be, to improve and to use information and learn as they go.
A corporate leader or a team leader or an organisation’s leader says, ‘What am I here for? I’m here to make the organisation better. How do I make it better? I learn from what I see.’
An organisation, a company, a country, is a learning system which is trying to improve all the time. Nothing there is new. What is new is that we know that there’s a lot more data to help us and more of it is being produced. We know there’re new tools and technologies that help us to do that and there’s now a lot of academic research which can be deployed to make sure that this vast amount of data can lead to improvements. But the hard part is actually making that relevant in different organisations rather than just being something that happens in a lab somewhere.
That’s where the disciplines from sport are super interesting. People are engaged by sport — why? Because sport is simple. I mean, everyone knows what the goal is, right? If the goal is to win, the metric might be goals in soccer or it might be seconds in F1, but everyone knows what the goal is and what the metric is. They know there’s an imperative to learn faster than the competition. So there’s a culture of experimentation.
There’s a culture of learning fast and failing fast because you don’t have infinite resources; if I have a new piece of information which might help me perform better, I know I will look at it because a marginal improvement is going to help me. But I also know that if it’s not going to help me, I’m going to move on to the next thing. So how do you build that continuously learning system — which is what a sports team is — in a world where there’s more and more data and there’re more and more tools for deploying that data?
How do you recreate that learning system in a much more complex environment like a bank or a pharma company? While on the surface there is a whole array of differences, there are also a lot of commonalities between them and we try to focus on those commonalities.
What are some of the blind spots in most organisations when it comes to data analytics?
It is super challenging because we all know that there are opportunities there, but very few know where to start, or who to partner with, or what technology choices to make. Especially — but not exclusively — if we have any kind of legacy. How do you know you’re not adding to your technical debt problem?
Where most people go wrong is in not thinking through, ‘What is value for us, and how do we measure it?’ Because if you don’t know that, it’s very difficult to be able to assess just how important this piece of the future is among many competing priorities.
The most common things we see are people looking around at the ingredients, and not starting with valuing an objective. What happens in many businesses is people say ‘I need a tool to solve this problem. I’m losing customers in a certain region or my product’s quality is poor — I want a tool to solve this problem.’ Or they might say ‘I’m gonna start by fixing all my data, I’m gonna get some data scientists.’
And we would say all of those things are important, but they all have to be seen in the context of what value means.
What are some of the ideal character traits or professional traits of good data scientists?
Obviously, we have excellent data scientists. But I see plenty of organisations with excellent data scientists who are not actually making much progress. I would say, the thing that is critical for us is the unitary whole of experts from all different backgrounds. We need talented data scientists, we also need talented engineers; most importantly, we need an environment where everybody’s work can be directed towards a common purpose. And they have the opportunity to develop their own skills and their own capabilities in line with our mission of making an impact.
That’s the thing I spend the most time worrying about: are we, as an organisation, as a team, maximising our impact and continuing our learning? And what do we need to do to learn better, faster, keep everybody motivated, and make sure that we stay relevant? Creating that ecosystem and that environment and that culture, and all the things that go with it, is the single most important thing. Then, of course, you need the right talented people to come and become part of it — the two mix together. We are absolutely fixated on talent, but I would say it’s not necessarily always — a team of geniuses without any balance or diversity and culture and organisation is, in my opinion, not going to achieve very much.
Sometimes, you don’t need to employ a genius to see the answer to a problem. We may look at a really complex problem — and once we understand it, talk to the owners and users, look at the data — say that an Excel sheet is a perfectly fine way of solving that problem.
A team which has a computer scientist, a data scientist, a data engineer, a designer…think about what those words are saying. What is a scientist doing? A scientist is experimenting, he’s trying new things. He’s saying, maybe we can try looking at it in a different way. What’s an engineer doing? An engineer’s saying, ‘I’m not interested in experimenting, this thing has to work. I am not going to put this thing out there until I know nobody’s going to die.’
The engineer and the scientist, in some ways, have a very different outlook on life. But if you put them together, it’s incredibly powerful. Then you have the designers saying, ‘Well, those things are all good, but if nobody wants to use them, they’re useless.’ So when we talk about interdisciplinary, we talk about ways of working, we talk about talent, we’re thinking very carefully. It’s what I mean by creating an environment and ways of working where you can make me have more impact than I would on my own. Otherwise, why would I bother to be part of this team? And that’s what we want everybody to feel. Be the best you can be, but also be humble and recognise that we’ll achieve more together than we would on our own. Otherwise, what’s the point?
How far away do you think we are from general AI that’s going to take over the world?
My short answer to that is that there’s a ton of work to be done with a human in the loop, which is going to continue for a long time.
And that’s what we’re really, really focused on at the moment. The singularity debate? God knows! We’re much more focused on — and we think there’s much more impact in — keeping the human in the loop and the power of human judgement. So augmented intelligence, if you like, is a more interesting concept for us in the real world than artificial super-intelligence.
