How Financial Times Built Data Capabilities Worth £3.2M (and counting!)

Tarush Aggarwal (00:00):
Welcome to another episode of People of Data, where we get to highlight the wonderful people of data. These are our data leaders from various industries, geographies, and disciplines. Discover how companies are using data, their achievements, and the challenges they face. Today on the show, we have McKinley Hayden, who's the Director of Data Value and Strategy at the Financial Times. Apart from spearheading data-driven decision-making and bringing data into the newsroom

‍

McKinley Hayden (00:19):
you

‍

Tarush Aggarwal (00:27):
She's particularly interested in the gap between theory and implementation and enjoys dealing with all matters, people, process, and culture. Welcome to the show, McKinley.

‍

McKinley Hayden (00:38):
Thank you so much for having me. It's such a joy to be

‍

Tarush Aggarwal (00:42):
I was super excited to chat. We were just having such a fun conversation before we started the show. So I'm really looking forward to this conversation. Are you ready to get started?

‍

McKinley Hayden (00:51):
I am, yes, let's go.

‍

Tarush Aggarwal (00:52):
All right, so question one, what does the Financial Times do? What role does data play in helping the Financial Times do what it does?

‍

McKinley Hayden (01:00):
So we are first and foremost a financially driven newspaper. We started in 1888, but I'd say our larger mission is around enabling ambitious people and companies to succeed. And so we actually have a wider group of products and services that not everyone knows about, but it's pretty impressive. So we've got, I think, 17 titles within our specialist titles, things like The Banker would be one of them. And we have also an FT Live company, which is all around events, FT Strategies, a consultancy that was born out of our success with our data and our digital transformation, which leads me very nicely to your second question, how does data help the FT and what role does it play? It's sort of everything. I had a bit of a light bulb moment actually when I went to a conference a little while ago where somebody was explaining how important it is to link a digital transformation with data investment and a sort of data transformation.

‍

And I was struck by this argument because I realized that we had sort of taken that for granted. We, as we invested in digital subscriptions and a digital transformation, we made an equal investment into our data capabilities. But I'm learning that plenty of companies have not made that link and therefore our...

‍

paying a price for it and having to do things retrospectively. I think we've just, it's kind of part of our history that data is a key engine for growth. John Ridding, our CEO, has been saying very publicly now that journalism and data are basically the two biggest competitive advantages that the company has. So I think that gives you a good idea of the regard for which data is held.

That doesn't mean, by the way, that we don't have work to do or that it's perfect, of course, but I'd say that almost every job around the organization is going to have some interplay with data and be using it for day-to-day decision-making.

‍

Tarush Aggarwal (03:32):
Amazing, excited to go deeper into how you guys are using it on a practical basis. But before we get into that, I know culture is particularly important to you. What would you say is the culture of the Financial Times, and how do you fit into that?

‍

McKinley Hayden (03:45):
So I think the thing that is so much fun about having new joiners at the FT is you get to ask them these questions around like, how does the FT culture meet your expectations? And typically people say a very similar thing, which is that they are so surprised at how warm and approachable and diverse the people are.

Because of our brand, sometimes there's a bit of a stuffiness that people might expect, but actually, the staff is made up of intensely curious people who by and large are pretty humble and actually are very just invested in everybody succeeding. And I think that that is a nice sort of garden, kind of, if you like, for a diverse and productive data culture, because if you don't have that essential curiosity and willingness to investigate and pursue excellence, like all the data in the world is not really going to help you. So I think we've got very kind of fertile ground. So yeah, I'd say on the flip side, sometimes we can be a little bit too nice.

You do need data sometimes to say when something isn't working. And I think sometimes we struggle. We're very happy to use data to say something has gone amazingly well, but we're not as good at saying, okay, this hasn't

‍

Tarush Aggarwal (05:28):
On that note, what does the data footprint look like? What is the team? What is the current data stack? Are you evaluating any new tools?

‍

McKinley Hayden (05:37):
Okay. Let me start. So I've been in the team for 10 years. And when I joined, our data and analytics department was like maybe 10 people, something like that. And we're now at about 70, um, which is, yeah, it's been quite a, uh, quite a journey. Um, so we have, um, we have five different teams. We've recently renamed some of them, so I'm going to have to refer to my notes to make sure I don't misrepresent. We've got data governance and architecture, business intelligence, which is in two different teams, engineering and reporting, insights and analytics, which covers strategic analysis, as well as experimentation and user and audience feedback.

‍

We then have data science and AI, which also has a couple groups, and then data value and strategy, which is my favorite team because it's mine. And so in terms of the tech stack, we use AWS to ingest our first and third-party data. We kind of push the enriched data into our BI layer through Airflow. Analysts typically use BigQuery to access the data there. But we've invested quite a bit into self-serve. So we have things like Looker, Amplitude, Bloomreach, Alida, Optimizedly, I think, would be some of the different tools that a combination of analysts and business users would be using.

‍

Tarush Aggarwal (07:20):
Yeah, beautiful. Any sort of upcoming areas for the business? Obviously, the question on everyone's mind is AI. Have you made any investments in terms of technology around that?

‍

McKinley Hayden (07:30):
What's AI? No, I'm joking. Yeah. God. No, yes, we've—we absolutely are.

‍

Tarush Aggarwal (07:31):
Haha. Great question.

‍

McKinley Hayden (07:43):
Like everybody grappling with the challenge. I think the thing that I am proud of is the fact that pretty early on the CEO recognized that this was gonna be a leveling technology and basically said, we've got to greet this in the same way that we did with digital subscriptions.

Our subscription model back in 2002 was seen as probably bizarre. And now it's even the Guardian's got their own model. Everyone, I think, Gen AI in particular is seen as another opportunity to be bold and greet what is an inevitable future where businesses who do not leverage to some degree will just be financially penalized for it. What I think is good about, so I think that's good point number one is to have that vision in the first place. The second one is that our first priority was to create an AI governance policy. So the potential risks of Gen AI are really substantial and I really...

I think to minimize them is does humanity a disservice? And particularly from a news organization that banks on its integrity and its quality. So after deciding to greet the technology with open arms, we invested time and resources into making a really robust AI governance policy that aligned with our company values and kept us in a relatively low-risk scenario whilst experimenting with these things. I mean, as you can appreciate, all it takes is one journalist making a foolish mistake and we could struggle to recover our credibility. So it's serious. So having done all that and kind of communicated that out, we've been using ChatGPT Enterprise and Gemini and a few other more kind of domain-specific tools. And we've also created an Ask FT, which is a Gen AI product for our FT Professional Sets, our B2B business, where instead of just coming in to search, you could type in Keir Starmer, for example, and now have this function where you could say, you know, what is Keir Starmer's latest policy around the European Union or, you know, whatever topic and you'll get this summary of what we have written about that particular topic. So that's kind of an early thing, but we're looking to do a lot more experimentation for sure.

‍

Tarush Aggarwal (10:47):
Very interesting. I think the governance aspect is something many data leaders are thinking about. On that note, what is one achievement which you're particularly proud of, an achievement where data has had a lasting impact on the business?

‍

McKinley Hayden (11:00):
Man, do you know what? That was the question that I sort of struggled with the most because there's so many. And something that has been talked about quite a lot is our North Star of RFE, so our Recency, Frequency, and Volume. I don't want to talk about it now because it's been done to death. So what I did instead was we've recently created a value methodology. So attaching a pound or pence, I was gonna say a dollar sign, but wrong country, some money to each of our data capabilities and assets and outputs. So I thought I would actually just use data to answer your question. So I looked across what we did last year and

My answer is we have, so the estimated value of our data capabilities and outputs totaled about 3.2 million pounds. So the most valuable category of output was our Looker dashboards, which contributed about just over a million pounds in productivity gains. But for a single initiative, the most valuable use of data last year was a digital splitter redesign experiment, which we were able, using our A/B testing capabilities, to show an option for the funnel that created about a 17% uplift in our conversion rate, which over about a year would increase

‍

Tarush Aggarwal (12:44):
That is immensely significant.

‍

McKinley Hayden (12:48):
Yeah, absolutely. That kind of— that should result in about 900,000 in average revenue per user and then 1.1 million in acquisition LTV. I mean, and that's a single test. So I'm going to use the data to then answer your question and stick with those

‍

Tarush Aggarwal (13:13):
I love the fact that you're using data to answer this. The last question we typically ask is what's, you know, the way I like asking this question is, you know, every company's got the ups and downs. What is one problem or one challenge which you're really excited about solving in the next, you know, six to 12 months and how are you going about

‍

McKinley Hayden (13:33):
So the biggest challenge that we have is essentially the dark side of a good thing. So I think at the FT, I've already described it as being kind of part of our bloodstream, Data is just a part of our culture and everything we do.

But funnily enough, it's when the downside of that is that it's sort of taken for granted. You know, I think data leaders are commonly frustrated by the, you know, it's just a magic thing, you know, like, it's just, just data just happens. And it's like, no, no, it doesn't just happen. And I think there's sort of an element of that at the FT that we have.

There is not as much understanding and visibility into what truly creates value from data. And that actually, if we want to create value from data at scale to meet our ambitions, then everyone needs to pitch in and it's going, and it's going to be uncomfortable. It is going to mean that some people are going to do less of the things that they want to do because there are other things that will need to be prioritized. So it's a bit of a twofold challenge.

Imagine having a best friend that's been your best friend for forever. And just every time you need, every time you just, you want somebody to go out for a drink with or see that movie that your partner won't go see, they're always like, yes, yes, yes. And imagine if one day that person is suddenly like, actually, no, we're gonna go see my movie.

Hopefully, you guys are good enough friends that you'd be able to make that work. But that's a transition, right? That's a change in the dynamic. And that's what we're looking to do. And I think we can do it. It's just, it is gonna mean it's a bit of a recalibration over the next year or so.

‍

Tarush Aggarwal (15:22):
What's important when you think about this sort of role, do you see this as a top-down initiative? How would you, on a practical basis? Because think a lot of data teams really struggle with that, right? And in some ways also, what we hear a lot is moving from this role of a help desk to really instilling kind of what you really lead, is your sort of brainchild, which is data value and strategy. So on a practical basis, given that this is something you wake up every day and think about,

What does the day-to-day implementation look like of this change in mindset?

‍

McKinley Hayden (15:56):
Yeah, I mean, I think so. I am a highly structured person and so I take a lot of refuge in frameworks and so on. And I know a lot of people who roll their eyes at that sort of thing, but I actually think when you're looking at this level of change and this level of complexity, it's really, really, really helpful to break things down. So I very much try to start with, you know, the top-end goal. So if we want data to be an enterprise endeavor, then that's the North Star, right? Then what's the next level down? What does that mean for people, processes, tools, and culture? And culture I define as, you know, beliefs, largely emotions, because these things matter, right? You know, culture eats strategy for breakfast, as we all know.

And then once you have those kind of four categories, you can map out like, okay, what are the existing things that are kind of keeping the current situation in enforcement? And how would those things need to be different to achieve that overarching goal? Then you start, then there you've already got like the semblance of a map, right? Around what are the interventions? And of course, you

on along that keeping kind of a list of your detractors or your ambivalence and, then your champions, you know, then again, you've got the people who can help you to enact some of those changes. But I'll tell you, it just involves a lot of talking to a lot of people and a lot of coffees and a lot of listening and a lot of, you know, and a lot of compromise.

And not compromise because obviously you've got to stick to your guns. But yeah, think having that kind of, I think if you don't have a structured approach, it very quickly becomes incredibly overwhelming.

‍

Tarush Aggarwal (17:54):
Awesome. That was a really great response. But, McKinley, thank you so much for the wisdom and thank you so much for being on the show today.

‍

McKinley Hayden (18:01):
My pleasure. Thank you so much for having me.