Multichannel Success Podcast Season 4 Episode 6 - Transcript

Becoming a data driven organisation - with Penelope Bellegarde

DW [00:00:12 - 00:00:52]

In today's episode, we're really pushing the boat out. And I'm delighted to be joined by a real thought leader in her space. In a previous life, she was a web analyst with the well-known travel brand, Chewy. She then joined PwC, boo, yes, as an analytics consultant. And more recently, she's become the founder of DataTouch, working to ensure leaders have the right actionable insights from data. She's also worked across a vast number of verticals, including banking, retail, media, FMCG, blockchain, automotive, the list goes on. Would you like to welcome DW? Welcome, Penelope.

PB [00:00:52 - 00:00:55]

Hello David. Hello Mark. Delighted to be here today.

DW [00:00:56 - 00:01:24]

I'm also joined, as usual, by my co-host, Mark, who this week has been exploring the wonderful world of e-bikes and why the hell they all end up in the same place at the same time. More on that later. But I'd like to start with Penelope on board around the subject of why this subject of creating a data-driven organisation is the right topic for today. Penelope, give us some sense of the imperative.

PB [00:01:24 - 00:02:22]

What I've been finding over the years is that everybody, every organisation is talking a lot about data and really wanting to become data-driven and sometimes thinking that they are data-driven. But in actual fact, that is not the case for most organisations that I see, whether they are big or small. And I do find that Everybody's hitting a brick wall in terms of taking advantage of data. Now we have AI coming into the equation and I worry that the same is going to happen. So it's very important to have some critical foundations in terms of, well, first of all, why do we want to become data-driven? Who should benefit? And what it is that we we really want to achieve? And let's write this down because data is not just a tech issue.

DW [00:02:23 - 00:02:39]

We were talking off-air about the beneficiaries of a good data-driven organisation, and I was delighted that your kind of purview of that is much broader than mine was. Tell us a little bit about who you think the beneficiaries of a good data-driven organisation are.

PB [00:02:40 - 00:03:05]

Absolutely. I think it should be everybody. Of course, it should be the organisation, but it should also be the wider ecosystem around the organisation, and by that I mean staff, employees, everybody should benefit from it, suppliers, customers, obviously, investors, and society as a whole.

DW [00:03:06 - 00:03:17]

which is quite a complex group of people with very differing and maybe in some cases quite polarised requirements. How do you manage the complexity of that?

MP [00:03:18 - 00:03:54]

I want to jump in there because I think from our perspective, and for the people that we come across who are predominantly consumer-facing organisations, that having been data-driven and being customer-driven often go hand-in-hand. In fact, I don't think you can be a customer, or as Forrester would call it, a customer-obsessed organisation unless you're also data- obsessed, because you've got to make sure that you are doing the right things, and the only way you can make sure you're doing the right things is by making sure that you're having the right data, and we will come on to the detail of that as we go through the conversation.

DW [00:03:54 - 00:04:18]

Now, I reckon we've done about four minutes of this and we've mentioned AI already. So, AI permeates a lot of this conversation. How advanced does your thinking about AI need to be? Because I suspect for the kind of clients we talk to, that's still quite a long way down the track. So, do you have to become competent in AI? Do you have to become advanced in AI? Or just need to know how to spell it?

PB [00:04:20 - 00:05:20]

I think what is crystal clear to understand from very early on for every organisation is, first of all, there cannot be AI without any data. So, yes, there are absolutely amazing advantages in starting to use AI straight away and experiment, but the foundations need to be there. So, the foundations behind data quality, for example, the foundations around privacy, data protection, ethics need to be there as well. And also, as we've just discussed, AI should be benefiting everybody. So, we also need to have that – to go through down the whole ecosystem in terms of what do we want to achieve with it. But I think the first component is, let's not forget that there is no AI without data.

DW [00:05:21 - 00:05:39]

You've led us quite nicely into a part we're going to come on to, I think, in a minute about the components of a good data strategy. But just before we do, that vision piece you talked about, Mark, I'm going to ask you, how important is having a vision? It's often something that's talked about, but it's handled very differently by different organisations.

MP [00:05:40 - 00:06:30]

We always find, it's the first question we ask a client, isn't it, is what is the vision, what are you trying to achieve? Because unless you have that, you can't hang everything else onto it. It doesn't matter whether it's data strategy or customer strategy or tech strategy, to be honest. It all has to hang off what is your overall organisation trying to do? Are you trying to be a low-cost provider? Are you trying to be high customer touch with great customer service? In which case you need a shed load of different data. That alignment at that level will then dictate some of the decisions you make lower down. Do you invest in building your own AI versus do you buy in a generic service? Do you invest in creating a data lake or do you just have an analytics package? Those decisions will stem from what you're overall trying to do.

DW [00:06:30 - 00:07:07]

I think that's a really good point. And if we need to have a comprehensive data strategy in order to be able to move to becoming a more data or customer- centric organisation, that presumably needs to play in line with the overall business strategy, the overall business vision. Because I think, you know, we've both had examples where the business has a vision of what it wants to be, and now it's got a data organisation structure, but there's no acknowledgement of that in the wider business strategy. So you end up with this kind of orphan child that doesn't really have any parents, it kind of has no support, it just exists in the corner.

MP [00:07:07 - 00:07:29]

If you think back to a client we worked on where we had long conversations with the investors about did they want to be a data driven business, a data enabled business or even a data agnostic business and we were having to explain that to them in order to try and get their guidance as to what they wanted and therefore that dictated how they were building their technology.

DW [00:07:29 - 00:08:34]

Yeah. And I think if you think about the kind of clients that we're offering, think about retailers and brands who often existed way before anybody told them that they needed to become a customer centric, hence data-driven organization, their vision clearly doesn't include those components because those visions predate. So there's sometimes a need for the new fresh way of thinking to permeate upwards into the organization, re-scoping, redefining its reason for being, which is kind of a difficult thing to do when the kind of cart is trying to push the horse uphill. Yeah. Right. Let's come on to components. I know it's a big part of this and we're big into giving our listeners some practical advice and some real world kind of examples of things to think about. So Penelope, I'm going to come to you. What are the big components that our listeners need to think about when they're in a frame of mind to become more data and customer led? What are the key components of developing a good data strategy?

PB [00:08:35 - 00:11:03]

So you've got a few to think about. In no particular order of importance, analytics and insights. What's the type of analysis that is being carried out at the moment? What value are you getting from it versus what you should be doing? That allows us to identify where the gaps are potentially and potentially everything is going well for this particular component. But that audit is worth doing because it always forces us to think about what else could we be doing? Okay, hang on a minute. Is it actually, is it going to add value straight away or do we need to park it for a longer period of time? So that's one bit. The other bit is the tech, obviously. But what I will say is that it's one bit in the list. It is not the biggest component of the equation. But this is one of the biggest issues I have been finding over the years is that for a lot of organizations, a data strategy equals a tech strategy or tech stack, or sometimes just having a data team. These are great and they are huge enablers, but they are just one piece of the equation. So again, same question with the tech. Okay, what tech do we have at the moment and where are the potential gaps? Let's put a plan together. Data point richness, to me that's something that's also a little bit neglected sometimes. What are the actual data points we have access to right now, internally and externally, versus what we should be having? How can we enrich our data ecosystem even further? And yes, sometimes it means by acquiring more tech, sometimes it just means by being a little bit more due diligent in terms of the data we already have and there's no need for additional tech. Because what I do find is a lot of organizations sometimes will say, oh we haven't got enough data, and actually when I go around I find there is way too much data available, a lot of it is not being used. But also what's even more frustrating is to see that the people working in the organization aren't even aware of that various pockets of data existing somewhere else. So it's all siloed. Exactly, exactly.

DW [00:11:03 - 00:11:26]

That's a really good point, I think, because I think in most people's minds, and maybe also including in mine, there would have been a temptation to believe that in order to be good at this, we needed more data. But actually what you're saying is maybe you've probably already got enough, but a lot of it's dormant and not being used for anything. So turn your mind to that rather than gathering more data from customers. Exactly. That you maybe equally can't use.

PB [00:11:26 - 00:11:59]

Yes. And there's a HBR, it's a few years old, but I think it's still valid today, survey, white paper, which literally unveiled all of this is that the problem that companies aren't able to, the reason why a lot of companies aren't able to see any ROI with huge big data investments is because they don't do enough with the data they already have. There is this race to more tech, more data. Hang on a second. Let's ask the right business questions first.

DW [00:11:59 - 00:12:11]

Lovely. I think we'll put a link in the transcripts to that HBR thing so people can read that. That's good. Any more components do we think before we move on to organizational change?

MP [00:12:11 - 00:12:56]

Well, I mean, yes, clearly, I mean, one of the things that we will both strongly believe in is that somebody has to have a measurement framework at the start of this whole process. I mean, we have seen that with clients big and small. If you think about a large grocery client, they did not have a measurement framework, and yet they had put in an expensive analytics package, which then was not tracking one of the fundamentals of analytics, which was overall site speed. How you can do behavioral data when you're not tracking site speed, which is probably the biggest determinant of behavioral activity on a website, I don't know. And that was all because they didn't have a measurement framework in the first place. So, for me, that's absolutely critical.

DW [00:12:56 - 00:13:03]

So let's just unpack that briefly. What is a measurement framework?

MP [00:13:03 - 00:13:32]

Essentially it is a, you tell me, but it is a document that structures the data that you're trying to collect and what you're going to do with that data, which then enables you to say, okay, this is important, therefore I do need to track it, even if it's slightly more difficult. I'm thinking viewed availability on a website, which is not that simple to track, but for a retailer is almost always worthwhile tracking. Right. Unless you have a measurement framework, you're not going to pick up on stuff like.

DW [00:13:32 - 00:13:41]

So this connects, if I'm right, this connects the theory of your KPI requirements to the reality of am I tracking that and therefore can I report on those KPIs?

MP [00:13:41 - 00:13:47]

Yes, I'm sure Penelope's more up to speed with measurement frameworks than I am.

PB [00:13:47 - 00:14:00]

I find measurement frameworks being so underrated and they underpin everything in the organisation that has anything to do with data remotely.

Mel [00:14:00 - 00:14:05]

We'll be right back.

PB [00:14:01 - 00:15:18]

Your KPIs are one element of the measurement frameworks and probably and arguably the most important piece. Once you've selected your KPIs and once you've selected the right KPIs, and we could do another podcast just on that, but you've got to make sure you've got owners accountability. So who is owning this KPI in the business? Who is going to get a call from the head of marketing, finance, content, whatever department it might be, when the KPI is traveling, sometimes in the right direction or in the wrong direction. This is, let's be honest, this is tricky because we are asking people, if you want to become data driven, this is what needs to happen. So people need to know that. Yes, you may have a little bit more on your job spec, you will become accountable. But the beauty of that is your team and yourself as an individual, you will become data driven, which is a highly rated skill in today's job market. And it doesn't have to be painful. It does require change,

MP [00:15:12 - 00:15:12]

Hmm.

PB [00:15:19 - 00:15:33]

and there are ways of doing it. But this is the only way to embed data in day to day operations, we've got to have an owner behind the KPI. And it can't just be everybody,

Mel [00:15:31 - 00:15:33]

And it can't just be.

PB [00:15:33 - 00:15:36]

which is the answer I sometimes get who's responsible for this.

MP [00:15:37 - 00:15:40]

You've got to have somebody's arm.

DW [00:15:37 - 00:15:45]

So now. Throat to choke, we used to say. Anyway.

Maybe we should take a break here for a message from our sponsor, Shopline, the first e-commerce choice for over 600,000 merchants globally. Shopline empowers merchants with the best solutions for online, offline and social commerce, from budding entrepreneurs to thriving global brands. So if you are looking to unlock the power of modern commerce, make it easier to manage and save money in the process, find out more at shopline.com.

DW [00:16:18 - 00:16:32]

Okay, so that's quite a comprehensive list of components, and I'm sure there are more, but one I just wanted to pick out which we haven't talked about yet is ESG. I know that's a favorite of yours, Penelope. Tell us a little bit about why ESG is a component.

PB [00:16:30 - 00:17:52]

Indeed, and I think it's, to me, it's one that is going to keep growing and growing, and also very degrading because tech tends to take such a big presence. So, environmental, social and governance issues, which a lot of companies are heavily being scrutinised against, and actually a lot of investments are going towards companies who can demonstrate that they are ESG driven. What does it mean with data? Quite a few things. On the environmental aspect, data centres, so where we send all our emails, all our photos, all our reports, where all of the machine learning and AI models are being built and run, that these sit on huge data centres, and these data centres require a huge amount of power and water to run efficiently. Some figures, and they are quite conservative figures, say that the electricity consumption of data centres at the moment is about 1-2%, it is very likely to rise, even though there are measures being done at the moment to try and mitigate this. So that's what you're saying,

MP [00:17:51 - 00:17:57]

So that's why people are putting in solar power data centres or Microsoft have immersed

PB [00:17:52 - 00:17:52]

that's what.

MP [00:17:57 - 00:18:07]

some of those in water so that they've got constant cooling to reduce the power output so they would be better from an ESG point of view. Indeed.

PB [00:18:07 - 00:18:33]

Exactly, which are really, really cool initiatives. But that is something that we need to think about both as data teams and also as citizens, you know, where is all of this data that I'm capturing day to day, at a fraction of what the cost was a few years ago, has a huge environmental cost. And one of the biggest stakes of calling the cloud the cloud is that.

DW [00:18:34 - 00:19:31]

I was just going to come on to that. I remember I had a conversation with my father. He said, it's up in the cloud. I said, no, it's in a data center around the corner. No, no, it's in the cloud. He didn't really understand what that meant, bless him. He's 95. Anyway. But no, it's a really significant point. And I guess probably most businesses kind of have to trust someone. They're unlikely to be having their own data centers. They're going to partner with someone who has. And therefore, how do they go about analyzing and making informed decisions about which partner to choose? Let's move on to, because you touched on it there, Penelope, a little bit, the organization and how the organization trusts or doesn't, as I think is probably the case often, that leading the organization to becoming more data-driven is the right thing to do. Because it doesn't necessarily, or maybe it does, touch everybody's lives. How do you kind of see that?

PB [00:19:32 - 00:19:53]

I personally believe it should touch everybody's lives. If we believe that data is a valuable resource, then it should. It should be used for wider purposes. And it can. It has the power of being used in that way.

DW [00:19:53 - 00:20:11]

So if I was the HR director, and I came from a slightly cynical perspective about this, I'd be saying, how does this impact me? How is this going to help me do a better job? How would you respond to that challenge?

PB [00:20:12 - 00:21:16]

I love that question because we didn't even talk about that particular example, but I've got a really good example from Google, who a couple of years ago wanted to solve one of the issues they had around new mothers leaving the business. And what could they be doing to improve that? And so what they did is, first of all, they ran an internal survey. Again, as a side note, I think that data is not just about quantitative information and numbers. They have their places, of course, but it's also about qualitative information, as Mark will, I'm sure, highly agree. And so what they did, they did this employee's internal survey around two things. What does a better boss look like for you? And what would make you stay more in the business? They did that, they acted on it, and they were able to reduce that rate of new mothers leaving the business. And so that is something that an HR department could highly benefit from.

DW [00:21:16 - 00:22:18]

Yeah, I mean I think there's organizations are always looking for efficiencies and they're always looking for technology to help them become more efficient and maybe that's a great example of that. I'm just thinking more in how you take those in leadership positions who are on the periphery of this, they may not be overly impacted by technology, they may not be overly impacted by the performance of the business, but they may still have significant roles and for some of them there may exist some cynicism about the value of becoming very data centric and I think one of the tactics I have used in the past is literally to talk to them and ask them to help me understand how I could help them do a better job. If data was put more at the heart of their business in their world, whatever their world might be, it might be you know I said it might be HR, it might be finance, how could better quality data delivered more effectively at the point of need help them do a better job and there are always answers to that.

MP [00:22:18 - 00:22:34]

Yeah, I would want to reference the podcast we did with Anna Berry, who spent many years at John Lewis, where they had cut back on their B&M; teams, because they're comparing themselves against Amazon at that point.

PB [00:22:27 - 00:22:28]

Yeah.

Mel [00:22:28 - 00:22:28]

Bye.

MP [00:22:34 - 00:23:00]

And they have suffered massively as a result of losing that human expertise. So it is a combination of the data plus the human interpretation of it that actually made them successful. which has caused some of the business problems they've been considering now.

DW [00:22:57 - 00:23:05]

Yeah, that's a really good example. Okay, Penelope, who should own data?

PB [00:23:04 - 00:24:12]

I know, it's quite a wide debate and you could understand why some people might say, well, it's the CFO, I own all of the data, I'm responsible for profits and turnover and costs, obviously. You could understand why someone would say, what is the CTO? Because data has been seen as just a tech bit. But we know the problems that come with that. Some organizations are recruiting for CDOs at the moment. The debate around is the appointment successful or not, it remains to be seen. But sometimes the CDO doesn't sit on the board. So there's also a bit of an issue there, I think. Sometimes it could also be a CMO responsibility. So everybody is playing some sort of key role around that. But let's see if we can make that a bit more transparent and see where there are areas of

Mel [00:24:08 - 00:24:08]

Sparrow.

PB [00:24:14 - 00:24:19]

of overlap and let's resolve that. Let's be open and transparent.

MP [00:24:19 - 00:24:41]

So do you think one role on the board should own the data, or do you think it works best if the CMO, the CFO, the CTO understand how data fits into their part, in which case the ownership is distributed? Or do you actually believe, what's the better outcome?

Mel [00:24:40 - 00:24:45]

We'll be right back.

MP [00:24:41 - 00:24:50]

Is it to have a single person in charge of data, or is it actually to make sure that that data is held?

PB [00:24:52 - 00:25:39]

I think it would be very difficult to have just one owner, even if that owner sits on the board. Okay. Because of all of the wider responsibilities that data involves. Again, it's not just a tech or a data point problem, it's also about what does it mean from a social and governance perspective, from a, I need to make sure that my staff, everybody's empowered to make decisions on the back of data, which is in a lot of cases is not the case. So I do think that it's more of a distributed and shared responsibility. However, I do think that it is at the end of the day, also part of the CEO's responsibility.

DW [00:25:40 - 00:26:30]

Well, I get that. I guess you could say that probably everything is part of the CEO's responsibility in terms of the organisational vision and strategy. But I come back to if a valuable component of this is having a clear vision, in the early days, someone has to own that vision and someone's got to drive that vision. I think it's like anything, when the organisation becomes mature at data gathering, data utilisation and all that, the rules can change. But to get yourself off the ground, to get yourself moving, to have a football team of owners rarely achieves the kind of result you want. I think you've got to start with someone who says, I'm going to own this, I'm going to drive it, I'm going to get us off the launch pad. And then once we're in flight, then we can have a healthy debate about exactly how we organise ourselves.

MP [00:26:30 - 00:26:55]

We've seen organizations who have a CTO, a Chief Transformation Officer, as their job title, who clearly, when you're trying to initiate the change and get things started, perhaps should be the single owner of these things. But then, once the business has matured, once it is perhaps a data-enabled or a data-driven organization, to some extent, then you get disseminated into the business.

DW [00:26:51 - 00:27:14]

It gets disseminated. Yeah, I agree. I agree. Okay. Look, guys, time is pressing. I think we've just got time for one more small segment here. How do I get started? What are the two or three things that I should make sure I am focusing on those things if I want to turn my organization into a data-driven organization?

MP [00:26:55 - 00:26:56]

Yeah, I agree.

PB [00:27:14 - 00:27:54]

I would start by asking yourselves better questions. Any team can do that. It doesn't require any data background, any coding, any math. Any team can do that. What's a better business question? It's a business question that is specific enough that is going to add value to your team or the organization and that is actionable. You will be able to do something with it because if you start with that, then it will become a lot clearer what data points should you be collecting. And then you can go to your data team and say, can you give me this and that, but asking better questions.

MP [00:27:55 - 00:28:52]

Yeah I would very much support that because we've seen lots of organisations who collect a massive amount of data but have no idea what to do with it. So I would challenge that and say make sure that you have a good understanding of what decisions you're capable of making on the back of the data and actually if you're not capable of making a change on the basis of knowing x, y and z why do you need to know x, y and z unless you have a long-term plan to get there and in this case you might need the historic data but that's a different point. So in terms of getting started making sure that you've got a measurement framework that understands these things that hangs off your overall business strategy, brings it all together, get an expert in to have a look whether that's you know somebody like Penelope, we can then help guide people to make sure that they are going down the right route to making the best use of data and that doesn't matter whether

PB [00:28:51 - 00:28:53]

And that doesn't matter.

MP [00:28:52 - 00:28:57]

it's somebody at a relatively crude level of understanding of data, you know, are you getting

PB [00:28:56 - 00:28:58]

Yeah.

MP [00:28:57 - 00:29:25]

the right market insight, qualitative data, all of that sort of stuff. Clearly, having the vast amount of data that you need to feed into AI, you know, you may have a very early conversation about AI and end up with a view that actually you don't have enough specific data from your organization or your industry to actually train an AI on. You'll have to use generic AI services from other people.

DW [00:29:25 - 00:29:40]

But be clear at the outset what you're trying to achieve by deploying AI because there I think you're into potentially significant investments and if you're not transparent and clear about what it is you're trying to achieve, that could be money fast approaching the plug.

MP [00:29:41 - 00:29:53]

If you're going to make your AI specific and helpful to your organisation, ideally it should be looking at the information that is specific to your organisation.

DW [00:29:53 - 00:30:18]

Excellent. Okay. Final thought from me on this would be, we talked at the top about the importance of having a vision. That is really important, but the best way to execute a vision successfully is to crawl, to walk, and then to run. Don't try and do everything in day one. As Mark was just saying, do it in steps that are appropriate to your organization. Okay, guys. At this point, I think we're going to have to wrap up because we've overrun. I'm going to thank you both to Penelope.

PB [00:30:19 - 00:30:20]

Thank you very

MP [00:30:20 - 00:30:22]

Thank you very much.

DW [00:30:20 - 00:30:22]

and to Mark.

MP [00:30:21 - 00:30:24]

Thank you, enjoy that.

DW [00:30:22 - 00:30:25]

Thank you for listening and we'll see you again on the next one.

Thanks to our sponsor Shopline, trusted by over 600,000 merchants worldwide to unlock the power of modern commerce. Find out more at shopline.com

Other similar Episodes: