Podcast Transcript of a discussion between Andy Davis, Head of Marketing, Venturi Recruitment and Mark Dodd, Head of Business and Co-Founder, Consolidata in October 2017.
Andy Davis (AD): From your perspective, what is that differentiator between a good data person and someone who is outstanding and really excels?
Mark Dodd (MD): I think from my perspective someone who is a data ’genius’ – someone who really gets data, is someone who understands the analytical techniques that are used in analysing and also understands what the data means to the business; actually, what is important to the business. A lot of data scientists and engineers they tend to like the analytics of the data but tend to miss the practical side of things.
AD: So, what would the practical examples be? I suppose that would fit a lot in with what you do on the business side; business outcomes and deliverables and how they fit in to the business purpose.
MD: Indeed. The whole aspect around data analytics really needs to fit into what the business is trying to get out of data and what their business strategy is. If it is a retail customer and they are trying to engage their customers to better build their brand, then they need to understand their customers, so therefore your analytics is going to be customer centric. Whereas if it a logistics business and you want to improve your efficiency – how full your vehicles are when you transport them around the country – then the metrics are going to be around volume and distance.
AD: So, is it about understanding those kinds of fundamentals? Does that make it more difficult to move from industry to industry? Because even though you understand the numbers, you have to understand the business outcomes. So, do feel there is value in people who go to particular industries and take to it quickly?
MD: A lot of verticals think that their data is special and their business problems are special, but I have been working in data for 25 years or so and I’m starting to come across the same types of problems in different industries. So, the advantage I have that I can bring to a problem, is that I have worked across 10 or 12 different verticals. There are some things like the reinsurance concept in the insurance industry that is fairly uncommon, but when you are engaging with customer data, product data or logistics data the concepts are very transportable across verticals.
AD: I realise you are more involved in the straight business side of things now than you had been before. Do you think that since you have become more involved in the business side that the way you approach data and the way you could transfer across industries has changed? You probably don’t give as much credit because you did it, or is it just out of interest?
MD: It’s probably a bit of both. For myself, I enjoy solving business problems that involve data and some technical solution. I work with a team of guys of varying skill sets, interests and backgrounds and together we can bring out good solutions. It is has evolved over time as everyone’s career does. My background is electrical engineering, starting work as a process engineer at one of Alcoa’s smelters then moved into statistical analysis. Over time it became more and more data analytics, not necessarily big data analytics but larger volume data analytics. We try to give something to our clients from the data that they didn’t know.
AD: So, from a client point of view, is that what they are after, trying to solve a particular problem, or trying to find something new?
MD: Interestingly, businesses are in many stages in their view on data. Some business still views IT and data as a cost centre; expensive to maintain and a problem to minimise; at the other end of the scale there are other businesses that see data as transformational. Their businesses are driven by the data – businesses like Amazon, WhatsApp – these are transformational data businesses. We meet potential clients that are only on their first DWH/BI implementation, even though other clients are on their third. A typical data warehouse life cycle is six to ten years I guess for a DWH implementation. Some businesses are 20 years behind others in terms of data analytics and BI.
AD: Is that something which poses different challenges from a client perspective? Which do you enjoy more? One you are building, the other you are optimising. Is it bizarre how little people are looking into data now?
MD: On some levels, it is bizarre how poorly engaged some businesses are with data. We prefer to come to greenfield or brownfield sites. Coming into an existing data warehouse that we have to make work when the original is not very well thought out and it was built five years earlier. If what they need to do is start again then it can be a difficult conversation. But the problem with coming into a business that is having its first data warehouse is that they will be very immature in terms of their understanding of data governance, stewardship, security; even data strategy.
AD: So, there is an education portion of the work. Because will it serve the purpose if we put in these fancy data centres and they don’t really understand what is going on?
MD: Indeed. It makes it a more difficult engagement in some ways. But we usually build long term engagement with our clients so they can get the best value from their data warehouse and BI systems.
AD: But still there is a lot of reporting from executive levels that they are not getting the ROI from those teams. It is not as simple that they are not doing their job. There are lots of nuances to it, that they were not getting what they hoped to from the investment, which again boils down to the education side of things. Also, there is the basic solving data problems with that intent. It is a problem for those organisations. Where are you seeing that the solution to the ROI type problem exists?
MD: I think the issue is that is a degree of marketing that happens with new technology. In the late 1990s and early-2000s, there was ‘BI systems are going to solve the business worlds problems’. From the mid-2000s to now, it is ‘big data solutions will solve those problems’. Now it has become ‘analytics solutions, data science, algorithmic systems will solve those problems’. With each of these generations of thinking these toolsets are all worthwhile, but it comes down to is the data strategy in line with the business strategy and do you know the business problem you are solving? Because to generate an ROI on a big data solution that is just recording tweets or mouse movements on your website has no direct ROI. You need to know what the business problem is. Are you trying to improve engagement and therefore how do you measure that? You then need to run an experiment, and say that the engagement for the last three months looks like this, we will change our website to look in this way and we will measure the change we see. The problem is that they don't express the business problem they are trying to solve and so they can’t measure what is the ROI.
AD: How do you solve that problem then? It’s almost down to a fundamental understanding of their own business problems and where the problems lie, without data in the first place are fundamentally not understanding them. Isn’t it kind of a catch 22 in some senses?
MD: It is, but I would say that the way the thinking should go, is that they should engage in a business strategy to decide what the 1-year, 3-year, 5-year business focus is. Then from the 1, 3, 5-year business strategy there should be a data strategy that also engages with the 1, 3, 5-year business strategy to support those business initiatives. If the business initiative is improving customer engagement, decreasing cost, improving customer margin, whatever those measures are, that will feed into the data strategy and therefore you can apply the appropriate toolkit to build data structures or to measure things or apply algorithms or analytics to support the business strategy.
AD: I’m not sure if this is the right way to say it, but for technical people now there is a lot more expected from them. I guess I mean around the business fundamentals and the understanding around that. It has evolved so much more than just the backroom staff or the ‘expense’ and things are changing quite rapidly now, so we are understanding that the technology is the product, the data will have a project ROI if we approach it correctly. With that, the business side are expecting the technical side to have a much stronger understanding of the business fundamentals even when you look at the high level technical role CTO. They are expected to have a really good technical understanding as well as understand the outcomes. I’ve even seen that filter down into more junior level roles; you have to understand the purpose of the technical work that you are going to be implementing. It is certainly an added pressure to technical people. Do you see it that way, that we are expecting a lot more of our technical people, or is it the inevitability of that technology is becoming so much more the product itself?
MD: Good question. I think necessarily technical people are no longer in the back room drinking coffee without access to the business and they are much more engaged with what the business is trying to achieve. But there will always be some technical people who are left with a ringfenced or focussed project that they don't necessarily need to engage with the business people that much, but there will be someone who can interpret what the technical person does and engage with the business. Because of the nature of IT people, I think a lot of us tend to be somewhere on the spectrum and those at the further end of the spectrum and perhaps are less people-orientated or people-engaged.
AD: It is tough, but from my perspective, we tend to expect a lot more from technical people, but from the business side, I don't know if they are expected to understand the technical. Maybe it makes sense for that not to be necessary.
MD: Just to go back to a comment you made a couple of minutes ago. I think the reason why business people don't engage too much in the technology is because the technology is fast changing and very complex. Although that said, there are a lot of very good end user tools of varying levels of complexity, Power BI from a Microsoft point of view, Tableau, QlikView and various add-ins to Excel that give business users a lot more power than they would have otherwise had. So, from that point of view, we are in a better place than we were 20 years ago. But there is little expectation that a business person would be able to engage with data, say through R, for instance.
AD: Are we talking data visualisation now and improvements of that. What we are talking about is how we can help them get involved more now?
MD: Yes, I think that is the main area that business is seeing paybacks from technology for themselves at the moment.
AD: Yes, it’s through the visualisation, right? I suppose that is right for anything, isn’t it? You wouldn’t expect them to dive into R and code away. Engaging with platforms like that to where it is cross boundaries; that’s the beauty of things like data visualisation, it’s all about that. Communication between a couple of teams.
MD: Yes. Certainly, the turnaround of doing visualisations in Power BI or more modern versions of Excel mean that business users can see something that is actually meaningful much quicker than they would have in say an SSRS report 15 or 20 years ago.
AD: I wanted to ask about your progression through data. I know that data has always been an interest of yours technically speaking. When did you zone in and focus on data as being the road that you wanted to go down? I always find from my experience as well, that there are data people are typically more aware of the business side of things; has that always been a constant for you; that you have always been compelled by the business much as the technical stuff?
MD: It has been. From my first job doing process control engineering, I did quite a bit of data analysis using SAS and I was able to engage with engineers in the business who wanted to make improvements in the process who needed data to support that and they would run experiments and through that application of real life statistics I saw that I could make a real difference by understanding and analysing data.
AD: When you started out and were working on different challenges what were the things that were spurring you on to continue to improve and continue to get better; a kind of hook for you for your career that you kind of had?
MD: For me it was that I felt that I could make a difference in the business I was in by digging into the data and looking at it in a different way and there weren’t many of us at the time that had the skills or the interest to do that and so the three or four of us in the business at the time were able to leverage this combination of skills and interest to be able to go and drive the business in terms of production efficiency.
AD: I can imagine that it still requires an element of creativity doesn’t it and 20 years ago, whatever data used to look like it was pretty bleak, whereas now it is much more of a creative form than when it was back then, so it was quite innovative to have to think creatively then.
MD: Indeed yes. It wasn’t quite that archaic, but certainly quite cumbersome. It was before spreadsheets were engaged with by business. It was analytics through SAS that was very advanced at the time and offered understanding and analysis that other languages were a generation behind almost.
AD: I’m conscious of time and don't want to keep you all day … insights, discovering things whether intentionally or unintentionally … unexpected insights from data … does that still happen so much now or is that something still as exciting if discover things by accident?
MD: One of our hooks that we have as a presales tool is we will ask our potential clients to give us some of their data, anonymised if appropriate, and we say to them, we will tell you something interesting about it, and it is likely to be something that you don't know. So, one of our now existing clients gave a set of data which showed twelve months of rental for a couple of hundred shops in a retail centre. We built a propensity model for these businesses for late payments, just from the data they gave to us – a bit of clustering analysis. OK they said, we never knew that was possible. So, there is always something interesting in the data, you just need to think a bit differently, sometimes bring in external data to enrich it and to build a bit of a story so that you can see where you can go, but there are always interesting things and a story in data.
AD: I like it. That is a good point to finish on, thanks very much.