datamaster summit 2020

9 Ways to Support a High-impact Analytics Environment

 

Nick Hall

Former lead for the Analytics Center of Excellence at Barclays

Watch a panel discussion and Q&A featuring Nick Hall, former lead for the Analytics Center of Excellence at Barclays. Andy Palmer, Co-Founder and CEO at Tamr, hosted this session to discuss how Nick Hall built high-performing analytics teams, unlocked value from customer data, and created new growth strategies that leveraged data as a strategic asset for the bank.

Transcript

Speaker 1:
Data Masters Summit 2020, presented by Tamr.

Louise Bolden:
Thank you for joining our session. I’m Louise Bolden, a solutions director at Tamr, and I’m delighted to introduce our discussion, “9 Ways to Support a high-impact Analytics Environment” with Nichol, the former lead of the Analytics Center of Excellence at Barclays. And co-founder and our CEO at Tamr, Andy Palmer, who will be leading the discussion.

Louise Bolden:
Today we have the chance to deep dive with Nick on lessons learned from building and leading analytics teams. Nick led the Barclay’s Analytics Center for Excellence for the last six years, to a team of over 140 analytics and data science professionals, who use analytics to grow the Barclay’s retail and banking business, as well as build data products derived in the bank’s transactional data. Since September, he’s been helping to grow a startup, not-for-profit. The Global Open Finance Center of Excellence, that will spun out of the University of Edinburgh in the first half of next year, delivering projects that leverage financial services, data for social, economic, and environmental goods.

Louise Bolden:
Nick also spent six years at IBM, and 16 years as a strategy consultant, building and deploying data capabilities. So Nick’s going to be joined by Andy. Andy is the co-founder and CEO of Tamr. During his career as an entrepreneur, Andy has served as a founding investor, a board member, and an advisor for more than 50 startup companies in technology, healthcare and life sciences. Andy was also the co-founder and founding CEO of Vertica Systems, a big data analytics company that was acquired by HP.

Louise Bolden:
All right, with that, let’s kick things off. I’m going to hand it over to Nick and Andy. So Nick’s going to kick things off with a short presentation, and then we’ll enter into some good discussion with Andy. So Nick, over to you.

Nick Hall:
Thanks Louise. Very good. Hello, everybody. Strange to be presenting, but that’s the new world we’re in. Let me start quickly with this slide. I don’t know what you guys did in your lockdown but myself, in common with maybe many people, decided to do a bit of clear out my home office. And I came across a white paper, which is certain on the left-hand side of this slide, which I actually wrote 20 years ago with some colleagues at a place called Mercer Management Consulting, which has since become Oliver Wyman. And it was entitled, “Making CRM Make Money.” It was back in the days when CRM was a buzzword in the same way that big data is today. So the essence of the paper was, we’ve got a lot of technology and CRM, but nobody’s actually making money. And the title was, “Technology alone won’t create value.”

Nick Hall:
Well, so fast forward 20 years, and I was giving another talk a few weeks ago, and I was using all these quotes on the right hand side of this slide from McKinsey and Gantner, and people like that, talking about analytics initiatives aren’t going to make money. They can’t scale, big data’s not going to make money. And it struck me that as the phrase goes, “The more things change, the more things stay the same.” And we have a lot of interesting and very helpful technology that is helping us. But every time we bring out some new technology, the doomsayers will tell us all the reasons why we’re not going to make money from it. So I thought I spent a little bit of time with this talk before Andy and I had a discussion. Thinking about what we might do from the people’s side in terms of creating our hand, I impact analytics environment.

Nick Hall:
So that can dovetail nicely with some of the really good, great, capabilities that technology can bring us. So, why do analytics teams struggle to have impact? I’m sure there are many versions of the answer to that question. Here are a few. Lots of valid reasons. Bad projects, bad resourcing, bad data. I’m sure Andy can tell us a bit about that, bad stakeholders. If you think it’d be so much easier if we didn’t actually have stakeholders that use our work.

Nick Hall:
They don’t use the work, they don’t make a decision, politics override the data. There’s lots of different reasons why these teams don’t necessarily get high value from what they’re doing. And the question is, what can we do about it? We can introduce some technological tools and things that help us, but is there something we can do systematically of the way we manage ourselves as an analytics team?

Nick Hall:
So I’m appealing to those of you out there who are either running analytics teams or who may be as part of your business, have an analytics function. And I just wanted to give a few ideas that we’ve had, through our experience at Barclays, around how you might support a high-impact, analytics team. And I’m an engineer at heart. It’s what I studied at university and it’s part of my philosophy. So I’m fascinated by how systems work. So I thought I’d throw together very rough and ready, a picture of the systems dynamics of my analytics team.

Nick Hall:
And it captures some of the trade-offs and pressures that we’re under and some of the ways that the team was working. So here’s if you like, a systems diagram of how the analytics team work. It starts with doing impactful work. So, one of the things I think you need to head for, if you want to have a valuable analytics team is, prioritize the most impactful work. Sounds pretty obvious, but there’s a starting point.

Nick Hall:
We’ve got to prioritize doing the stuff we think is most impactful. And of course this was in a private sector. This is a bank where I worked but it’s true of a public sector environment, it’s even true of social impact, anywhere where you’re trying to use analytics to have a positive impact on some aspect of your organization. You want to have impact, and that impact might be profit. It might be revenue growth, it might be cost reduction, might be capital savings. It might be improvement of customer experience, or it might be citizen wellbeing, or it might be some form of social impact. But the point is, what we want to do is, we want to have impact. So, part of doing that means we better measure the impact that we’ve created. And so measurement, pretty obviously, you don’t get it if you don’t measure it.

Nick Hall:
You realize what’s the impact you’ve had. And I think one of the things you can do with that if you’re within an organization is, you can use that amongst other things to grow your team’s reputation for doing high-impact work, right? So, if I’ve done good work, I can demonstrate the impact I’ve achieved, and I can use that to grow the reputation and why you want to do that? Well, because that means you can then take on more higher, impactful projects, and that leads you around a nice little virtual search. It’s a nice search, still going on here around. Find the most impactful work, measure what you’ve done, use that to talk about what you’ve done, and then use that to take on more higher-impact projects.

Nick Hall:
One of my colleagues once described this as the pie-eating competition, where first prize is more pie, but that’s exactly the business me and my team were in. But it goes further than this. This is far too simple for a good systems, dynamic diagram, but if you measure the impact of what you realize, you’re going to learn fast about what works. So one of the major benefits of measuring good impact is learning fast about what has worked. And if you do that, one of the interesting things you should do and can do, is drive up the productivity of your team, because you know some of the outcomes that are effective, you know what analytics leads to good outcomes. You simply learn what gets adopted and what gets used in the organization you’re part of.

Nick Hall:
Whatever it might be, learning fast allows you to drive productivity. And if you drive productivity by definition, the capacity of your team goes up, and if the capacity of the team goes up, well, that means you can take on more higher-impact projects and so feed that circle. There’s another benefit of measuring impact, which is, if we think about some of the techniques that are coming along the line in data science, in some of the technology that’s available, in some of the new tools that are available, if you really understand what’s working, that gives you a much better insight into how can I raise the sophistication of my analytics? What’s working? Where are the gaps? Where might I have more impact with better tools or better methodology? And actually, it’s only when you really have a good understanding of what projects are working and why in your analytics environment. Do I think it really makes sense to start? Let’s throw some advanced data science at it, let’s throw some advanced analytics at it.

Nick Hall:
Really, you can only start to apply that sophistication once you have a good understanding of what’s working. So you can raise that sophistication of that analytics. What does that do? It feeds the team’s reputation, because now you have a reputation. Not just doing impactful work, but actually you’re even smarter than people thought you were. So, that grows your reputation.

Nick Hall:
If you grow your reputation, and here’s I think perhaps, the kicker. Growing your reputation and having great impact allows you certainly in a business environment, to argue for investment in more people and more tools. So if you invest in more people, and more productivity tools, grows your team capacity and boom! Take on more and higher-impact projects. So, a simple little diagram, well not that simple now, but a simple little systems dynamic diagram that shows you two or three feedback loops that are going on I think, in the team that I was operating. And I thought we would focus on, we’ve only got 20 minutes. I thought we’d focus on three areas here and just go do a little bit of deep dive into three things.

Nick Hall:
So one is this idea of driving productivity. One is this idea of prioritizing the most impactful work, and the third is around learning faster. Those are the three areas where there’s quite a lot of richness in terms of some of the things we were doing. And I thought you’d all find that quite interesting. So in true BuzzFeed style, because we have 20 minutes so we have to do this at pace. It’s always 21 ways of doing X or 27 ways of doing Y. This is nine ways to drive your analytics team to shoot high-impact results.

Nick Hall:
So we’re just going to blow through those three things. And for each of the things we’re going to do a bit of a subdive into three areas of each of those as aspects. So let’s start with prioritization for impact. So the picture on the left of this slide is highly disguised data, but it is a fair representation of the way we would look at our team’s activities. So we would divide up into different areas of the business and we would measure as best we could, the PBT impact we were affecting. And we would track that with some emotional charge outright. So imagine we were charging ourselves out as a consulting service on the outside, what might we charge? We didn’t actually charge that number, but the idea was we would measure essentially our costs to the business. And obviously what you’re trying to do is achieve a return on that cost. And so, we did two or three things to help us manage that picture on the left.

Nick Hall:
First one, envision high-impact projects. What I mean by envision is shifting the balance of the team from a reactive. The business tells you, I want to do this analysis to us actually going proactively out to the business with ideas for ways we can have impact on the business. So it’s much more of a proactive scoping and design process. And to do that, one of the things you need to do is have pretty good trusted relationships with your stakeholders in the business. So we spent quite a bit of time actually thinking about how do you build those trusted relationships?

Nick Hall:
I think the other thing you need to do is for a project to be high-impact, you’ve got to take the project to the point at which you either make a decision, or the business makes a decision, or somebody actually does something to make an investment, or they take a decision, or they decide to go and activate a certain segment of customers. But actually thinking about projects in terms of all the way to that point of impact was an important step for us. And as you get good at deciding these are the good projects that we should be involved with because these are high-impact projects, actually going out and pitching for those opportunities, almost as if you were an external consulting firm, rather than simply waiting for them to come in and audit it. But actually going out and saying, “These are the things we ought to be doing, which are aligned with the strategy and which we think analytics could help you with.”

Nick Hall:
Second thing is track the impact as it talked about, and so getting pictures like the picture on the left, where you look at the value by the project, you start to package stories around. Here was the impact we have. We even ran a notional shadow P&L if you like, that talked about, this is the value we’ve driven to different parts of the business. And then actively managing that project portfolio. Particularly, the old fail-fast thing. So where you see projects declining in terms of the impact that they’re giving, or in terms of the insight you’re getting. Killing those fast, so that you can use that capacity elsewhere. And then the flip side of that, is scaling fast what works. So, there’s a big 80-20 here, and 80% of the value gets delivered in the first 20% of the work. So managing that dynamic and saying, at this point, we’ve had a lot of high impact, now it’s time to move on. And being quite agile around how we manage projects up to a MVP. Seeing if that has impact, and then only scaling up and rolling out when we see impact has happened.

Nick Hall:
That again was a really good, really critical skill for managing that value and prioritizing for impact. So, that’s prioritization. Second thing, learning fast. And I guess there’s a micro and a macro story here. From a macro perspective, raising the quality of everybody’s skills in the team was something we spent quite a bit of time thinking about. And I would start that with, any analytics team that is going to have good impact has to have great domain knowledge of the environment it’s working in.

Nick Hall:
If you’re in banking and you want to work with the lending business, you’ve really got to understand the lending business. You want to work with the mortgage business, you’ve really got to stand mortgages. So I’m actually actively building that knowledge into the team with something we spent time doing. If I was to say domain knowledge, I think the other big piece that I would say characterizes a team that really has impact, is one who thinks about problems in an end-to-end way. So what I mean by that is, it’s back to this idea of thinking through the problem until an action gets taken, until the decision being made. So solving problems into end-to-end way, which often means thinking a bit like an owner. So if I’m going to do work for the current account business, I’ve got to think like I owned the current account business, and think about what’s important to the person running that business and how can analytics help them.

Nick Hall:
So I think teaching that approach or thinking like an owner and thinking about problems end-to-end is something we thought a lot about. And how do you inculcate those skills? You can’t do it from a classroom training or go read a book. We actually built in a bit of a coaching mindset into the team. So part of what we were doing is running an apprenticeship environment, where the more experienced people would coach the less experienced people. And we even codified this into an academy we’d run each year. So some of these basic foundational skills that everybody needed to know, we would package up and run in some form of academy, either through the year, or in certain big splashes during the year.

Nick Hall:
So that was raise the floor if you like, get everybody’s skill level up. And the other side of this is raise [inaudible 00:16:50], raise a ceiling of certain individuals. So, I think you’ve got to have a systematic program of taking some of the more advanced techniques and saying, how can I see those into the business? Both in terms of the skills of the individuals and in terms of the projects where they get used. So obviously, supporting continuous professional development rarely talks about continuous professional development being a good thing, but from an analytics perspective, I think managing that at a team level can be really valuable, because you start to think these are the skills areas I need more of, and you can start to encourage certain individuals to develop their technical skills more in those areas. This tends to be much more in the technical side of the skill sets.

Nick Hall:
So raise the floor, raise the city, that’s the macro use of learning. And then from a micro perspective, I think in common with many teams out there doing optimization analytics, we did a lot of robust testing. We had a bit of an experimentation mindset. If you’re going to do something, can you do it and have a control so you can see what’s worked and what’s not, and measure mirror against the baseline properly?

Nick Hall:
So I think that was just a cultural mindset that is absolutely critical. And then, on a more informal basis, having forums where people could share stories about projects, and then having maybe a structure, like some knowledge management tools where you could actually start to share projects that work, and stories about projects that work, and certain methodologies that work, was incredibly valuable. And the last one, driving productivity. Maybe we can segue a bit into the conversation with Andy, but certainly driving productivity of the analytics team is really the best route you have for driving up capacity in the short-term. And I have the three things that we thought about quite specifically here.

Nick Hall:
The first one is, leverage your best. It’s not always that you’re sometimes capacity constrained. Sometimes you’re just constrained because there are two or three individuals that are always in demand. So how do you get over that? One of the ways is creating almost pyramids of teams within the analytics structure, where you’ve got your very best and your brightest problem solvers being leveraged by those that are learning. So again, it’s best inside of an apprenticeship environment, create good leverage around your very best analysts. And one of my favorite words, we used to talk about fungible people. And a fungible person, is somebody who’s got that raw, problem solving skill set, and can move amongst domains and apply that knowledge in different domains in a way that makes them easy to move where the need is.

Nick Hall:
And this obviously drives your productivity because it allows you to move capacity fast from project to project, and allows you to take people from lower impact work and move into higher impact work. So fungible people, are your friend. The second thing that we did here was, support those analysts with some of the things that might get in the way, remove some of the barriers that get in their way. And in many environments, I’m sure you’ll find that analysts seem to have stuff that gets in their way. It being, I can’t get the data. I got to do a lot of data cleaning and my environment isn’t working. I need to liaise with tech, I need to liaise with ops, I need to get stuff done in order to create the space where I can do the analytics that I’m paid to do.

Nick Hall:
So we ended up actually putting a ring fencing, a small team of people who did this day in, day out. So they really understood our environment. They really understood the infrastructure. It was almost an ops team within the analytics team. And these people both supported the tool sets of the analytics team, and were skilled in thinking about, “How do I remove barriers to doing good work?” And the third thing is, finding tools that focus on analytics productivity. And I deliberately mean the productivity, the analysts, as opposed to general productivity. And there’s a lot of tech out there, and certainly, from a technology and tools perspective, we looked at, and considered, and tested, and worked with a number of different tools that helped us in our environment.

Nick Hall:
And there are a lot of tools nowadays, which are rightly trying to democratize data and enable people to do analytics, even if they’re not trained analysts. And those can help an analytics team have high impact, if they remove some of the low value-adding work from the team. But the tools we were really interested in, were ones that helped improve the productivity of the analytics population. So really focused on the productivity of the best analysts. How can we do it to remove about low value-added work, help with data quality issues, help the team get productive fast, ingest data fast, and get to the point where they’re doing analytics faster? And so, those items, leveraging the best, supporting the analysts, focusing tools on analytics productivity. Well, three of the things we did to drive our productivity as a team.

Nick Hall:
So there are nine things, hopefully those made sense. Of course normally, if we were going on for another hour, then the real question is, well, how do you do all this? That’s the “what.” So what’s the “how”? Sadly, I haven’t got that much time, so I’m going to skip that question and to leave with this slide. And one of the answers is, the culture of the team. I think a lot of those things we have embedded in our culture and as Drucker said, “Culture eats strategy for breakfast.” It’s clearly the most powerful way you can get a high-performing team to adopt different best practices and different practices. So that’s my answer to the how, but the real answer is… afraid of the time. That’s it. I hope that was helpful. Andy, over to you, or rather over to us, I guess.

Andy Palmer:
Nick, really incredible talk and so much content in such a short period of time. John Shank would be very proud.

Nick Hall:
You’re gonna have to explain yourself to him.

Andy Palmer:
Yes, absolutely. Yeah. Nick and I had the opportunity to go to business school together over 25 years ago. So, that lots of shared history.

Nick Hall:
Yeah. We have history.

Andy Palmer:
And amazing that we both ended up in the same spot, working on data and analytics in large companies. And so many of your principles are consistent with the things that we believe at Tamr, and that I’ve experienced in my career.

Nick Hall:
Yeah.

Andy Palmer:
And I’d love to kick off the conversation, first by talking about this idea of people being at the core of the challenge, and that the people’s behavior around their analytics and how they think about prosecuting these things is so compelling that all the technology and tools exist, and analytics tools has been democratized. Everybody’s got visualization tools, and modeling tools, and these things, but how do we get people to actually organize around turning those tools into real value?

Andy Palmer:
And one of the things that really struck me is this idea that you have to encourage people to learn fast what works and what doesn’t, and then adjust. It seems a lot of the projects we get involved in, people get caught up in process and dogma and want to just keep the same thing over and over again, is antithetical to what you described. So maybe tell me a little bit about the experiences that you’ve had doing it, and what are the things that you’ve seen that have failed that are… Are there patterns there?

Nick Hall:
I mean, of course there are two levels of people, there are more than two levels I’m sure, but there are two levels of people here we’re talking about, which is, and I would cruelly separate them into us in the analytics team, and then the stakeholders. Of course, it’s never that clean and hopefully, you’re working very closely with the business. But I do think there’s getting your team to be data-driven, which most of the markers that’s their natural profession. But then the real challenge is how do you get both the business to be data-driven? And I think there were a couple of times when we actually did work where we were almost deliberately helping educate the business, and that can be very effective.

Nick Hall:
You have to be given permission to do that. You want to go out and say, we’re going to teach you all to be data-driven. That’ll soon get you kicked into touch. So you’ve got to be careful about how you come across. I was telling a story a couple of weeks back actually about when we first introduced this concept of test and control. So the concept of testing is a huge cultural barrier for many businesses, because the idea that I’ve been 20 years in this business, I know the right product when I see it. And I know the right answer when I see it. To say then, that you should surrender yourself to rigorous testing, and go with the strategy that the test shows work can be a real cultural hurdle.

Nick Hall:
So one of the things we did was, we all deliberately built a project around building and testing into a part of the business, with a little bit of show-and-tell and executive teaching around the value of testing, why test and control works. We talked about the concept of the God complex, where people think they know the answer, but they don’t. And you can use lots of great examples from many good books written. Great examples from medicine that talks about why rigorous experimentation is the gold standard of analytics. So I think one answer to your question is this idea of maybe putting exec education, maybe building that in as a component of how you feed back results of the project. But the second thing is a bit of humility from the analytics team, I think, which is, the data can never have the answer, right?

Nick Hall:
In fact, there’s almost an interesting conundrum where you can say, always trust the data, and never trust the data. And by always trust the data, I mean, always ask, what data can I use to help me with this decision? But always distrust the data in that, data’s always flawed, it’s biased. It only shows one part of the problem. If you’re doing something simple like, where do I open or close or refer Bishop branch in a branch network? I mean, there are a dozen different bits of data you can bring to answer that question. Then there’re also dozen qualitative things around who else is in the town? Are we the last branch? We are the newest branch.

Nick Hall:
What’s the nature of the local area? So, there are so many nuanced management issues. So in that situation, part of what your job is as an analytics team is say, “Look, here’s five or six criteria that lend themselves to data. I’ve done the analysis.” On this five or six criteria that don’t lend themselves to data, which is a commercial judgment, it’s your job as a management team to put those things together and make the decision. So I think you’re both bringing the data to bear, and you’re helping facilitate them to make a choice. But the really important thing is having the analyst’s team around when they make a choice. I think if you’re in the conversation, when they’re making a choice, that helps you influence whether or not data is brought to bear.

Andy Palmer:
What’s amazing, you said, because it feels like you’re in training with them, right? You’re building up this muscle to use data and analytics to help make better decisions, and it’s not a binary thing. It doesn’t happen overnight, and role models probably help a lot. I mean, you guys must have had some early wins that you used as role models, right?

Nick Hall:
Yeah. So, I’ve never worked with Google or Netflix, so I would imagine, and I don’t know their management teams, but I would imagine that the environments are a little different and maybe the whole of the management team is more data-orientated than your typical, larger, medium-sized company. But exactly to your point, you were using examples of where it’s worked, and you were using examples of stories of… And literally, almost saying we saved money here, or we didn’t. I’ve even used stores where we demonstrated, this makes you very popular. We’ve demonstrated that a decision had it gone differently, had they used the data, they would have saved money. So you can use both this carrot and stick here to say data would have helped. I prefer to show the success story.

Andy Palmer:
That’s great. Well, switching over to the data side of the equation, which is always our thing at Tamr…

Nick Hall:
Yup, It’s really.

Andy Palmer:
One of my mentors was this guy, Marvin Minsky, who always said, “No algorithm is useful without enough, great data.” But when you think about a place like Barclays, which has, I’m sure, incredible number of systems and lots of silos of data, how was that a part of the challenge that you had at Barclays? And how do you think that extrapolates out to what other companies are going through in terms of bringing data together and doing the analytics successfully?

Nick Hall:
Yes, we had huge amounts of data, and you always find yourself wanting more. I mean, I think one of the big challenges that the bank was really getting its arms around, as I left, I think it had been for the last couple of years, was this. I think it had gone through a period where, let me see. I think this is something very familiar to you. So the data sits in silos, right? And the data has been optimized for running a particular part of the business. So for example, if it’s the system that opens a current account that’s used by the teller in the branch, there’s a whole set of data around how to run that operation.

Nick Hall:
And one aspect of getting a bit more sophisticated about how you use data, is you start to do things like, “Good heavens, we’re not just a branch network, we’ve got telephony, we’ve got digital, we’ve got mobile. Let’s start to think about the omni-channel network.” So let’s start thinking about channels. Well, so that suddenly means to say, I’ve now got to say, I’ve now got to start to measure the aspect of common things that you’re doing in the branch channel with the same activity in that digital channel, with the same activity you’re doing in the telephony channel. And those systems don’t track that data in the same way, right? So, opening the account indication in one system is completely different from the opening the account. So, we were dealing with that a lot in some of the analytics we were doing. Trying to do that horizontal analytics across the business, across different businesses, across different channels, across different customer segments. And you’re trying to knit together data from operational systems that was never intended to be used that way, but you want to do that. You can’t build a system from scratch.

Andy Palmer:
Yeah. And I love that idea you had, where as an analyst, you have to think as an owner of the business, think more holistically. But then the underlying data is organized in a way that’s inherently idiosyncratic to any one group. Resolving that is something that we struggle with a lot with many of our customers. You have these analysts that want to think like owners and more holistically, but they’re overwhelmed by the difficulty of bringing the data together at that scale.

Nick Hall:
And also definitions. And I think I’ve heard you talk about this before, but this idea of… I used to laugh. Here’s an easy question, how many customers you’ve got? For most of us, how many customers you’ve got? Should be an easy question, right? But of course, if you’re an owner of the business and particularly, let’s say I’m running a current account business, and I want to say, how many customers are doing X? I probably don’t actually want to know exactly how many customers. I want to know how many active customers, because that’s probably what’s important, because I’m probably thinking about something that I want to do to that group of customers, or introduced to them, or I’m worried about a change I might make. So I’m really interested in people that are very active with me. Well, I mean, that might only be a fraction of the base. Sometimes it’s a surprisingly small fraction. Sometimes it’s a large fraction.

Nick Hall:
But the point is, somebody’s got to agree what that definition is. I mean, there are multiple ways you could define what an active customer is. So, you’ve actually got to get to a point where the management team, who may not be as interested as you and I are in the intricacies of the data, you’ve actually got them to actually decide. You guys must pick. Is an active customer somebody who’s been inactive for the last three months? Is an active customer somebody who’s done two transactions the last six months? It could be any of those, but the point is, you need to agree. And I think that was one of the big things I wouldn’t say we struggled with, but it was just… If you ever see two teams in the business doing analytics and disagreeing on the answer, nine times out of 10, both teams had done good work. They’ve just had different definitions of what they were looking at.

Andy Palmer:
It’s amazing you say that. We run into this a lot when it comes to classification, right. I just say, this is a classification problem, right? It’s an active customer or not. And we ran into this quite a bit when it came to parts classification and spend optimization when you get people describing the same simple little thing in two very different ways.

Nick Hall:
Yeah.

Andy Palmer:
We believe they’re right, and in some ways, we advocate at Tamr, for an approach that says, “Well, in each one of your systems, you can call it what you like, but when we look at it overall, we’re actually going to call it something that’s consistent.” And making the mapping to the idiosyncrasy in each system.

Nick Hall:
Presumably in large corporations, what you want is, the analyst that’s on the other side of the world could get access to that definition and know what it is as opposed to…

Andy Palmer:
Yeah. Exactly. And especially when you’re doing traditional spend optimization, invaluable. But it’s like there’s this respect for what people call things in their own systems, in their own little world. It might be okay, right? If you can get the best of both, have an integrated view, and also let them keep their business running the way they do with the systems they do, and the way they call things. That’s the best of both worlds. And it’s possible to do now with some of these next-gen techniques, like the stuff we do at Tamr and avoid.

Andy Palmer:
I don’t know if you see this, but one of the other topics I’d love to talk about, standardization, right? Oftentimes, people think about the way to solve these problems is standardize. Well, we just have to standardize everything. And when I was a CIO, I had this thing, you just never got there. Right. And the systems never actually became standardized, and the projects to do the standardization took a decade. And so you embrace the idea you’re going to live with heterogeneity.

Nick Hall:
I think you have to. I think it’s the old XKCD cartoon where there are 10 power adapters, and then we must have another power adapter. So we bring an 11th power adapter and in a years time, there are 11 power adapters out there. So each wave of standardization, you’re right, just creates another variation as opposed to often rather than the crane standardization.

Andy Palmer:
Yeah. Well, it’s amazing. We all see these things the same way having practiced in the real world. And I think that one of the other things you said that really was inspirational to me was this idea of the humility of the analyst in the process of prosecuting these opportunities. I think it’s very, very powerful, and maybe a real contrast to the likes of, I mean, you were a consultant for many years. The consultants come in from the outside with a lot of hubris, and if part of what we’re doing maybe is building this analytic muscle inside of these companies so that they can do a lot of the things that they would have hired Mercer to do. That building some humility into that, it’s a real differentiator from hiring the overly confident external analysts.

Nick Hall:
Absolutely. Well, but it’s funny you say that because I think the humility and that be humble, build trust relationships with the business, was something I took from my consulting experience. And I think the consultant has to be arrogant in a way, but I mean, only arrogant in the sense of, I know some things that could be valuable to you. But if the consultant from the outside doesn’t build trust and an advisory trust relationship with the client, they weren’t going to work. Since then, that was one of the skills I think we did develop quite a good muscle around, building an advisory relationship, which means, for example, having a shared agenda with whoever you’re working with in the business.

Nick Hall:
And that’s not always easy, because there are politics in business. And in fact, the flip side is sometimes, internal analytics teams can be quite arrogant because if you’ve been given a mandate that you do all the analytics, then there might be a tendency. I’m not saying this is, but there might be a tendency to say, “Well, we’re the Alix guys, and we say, this is X.” So, that’s the answer you’re going to get. And equally, you might also say, “Well, I work for the CFO and believe me, the finance agreed answer is X. So take it or leave it.” Whereas I think a really effective team, who is genuinely being an analytic function with an organization, needs to be able to work for the chief marketing officer, and do something for them and recognize that they have a certain agenda and you need to tread that delicate line of, “Well, I’m not going to do analytics wrong, but I’ve got to realize this guy has a marketing agenda that I’m helping with, and that’s what he’s doing the analytics to support.”

Nick Hall:
And I might have a different agenda from say, head of product, and I think teaching people nuanced ways of managing that is quite important. I used to like to say, “We don’t play politics, but we understand politics.” I don’t think you can survive in a corporation without understanding the politics going on. You choose whether or not you can play those politics yourself. I like to think we didn’t.

Andy Palmer:
It’s really remarkable. And these subtle behaviors you’ve described, I think can be the difference between winning and losing when it comes to building this muscle and truly remarkable, everything you’ve accomplished and your principles are so strong. We at Tamr, we’re your biggest fans.

Andy Palmer:
The things you espouse and the work you do makes things so much easier for us because we just want to clean up the data so that people can put it to good use.

Nick Hall:
Absolutely. Yeah. There’s no point in cleaning up that data if people aren’t going to use it. So, I can believe that you can see those philosophies are valuable and very good.

Andy Palmer:
It’s really been fantastic to catch up and really pretty good time and look forward to once this COVID craziness is over, getting together in London and sharing a beer.

Nick Hall:
Or Boston. I might get to Boston before you get to London, who knows? I miss it terribly.

Andy Palmer:
So either way we get. Thanks Nick.

Nick Hall:
Excellent. Good to talk to you Andy.

Andy Palmer:
Cheers Nick. Bye.