<img height="1" width="1" style="display:none;" alt="" src="https://dc.ads.linkedin.com/collect/?pid=306561&amp;fmt=gif">

Takeaways from Corinium's Chief Data and Anayltics Officer (CDAO) Sydney Conference

Written by Corinium

Takeaways from Corinium's Chief Data and Anayltics Officer (CDAO) Sydney Conference

Written by Corinium on Jul 6, 2018 10:39:58 AM

News Artificial Intelligence & Machine Learning Articles

The Corinium Chief Data and Analytics Officer (CDAO) conference in Sydney (6-8 March 2017) was one of the better conferences I’ve attended recently. I was up till midnight most days clearing BAU work so that I could attend. I’m glad I made the time for it. Great keynote from Mark Hunter CDO of Sainsbury’s Bank. Overall, lots to digest from data wrangling, feature engineering and dimensionality reduction techniques to agile methodologies, tribes and scrums! My favourite quote from the conference “All Science is data science … the reverse is not true” from Teradata, Data and Ideas Therapist Dr Clement Fredembach. He’s the guy that looks like a mad scientist.

As a CDO of a start-up with a team I can count on one hand, I felt somewhat under represented relative to the larger corporates in attendance. Although exhibiting vendors were courting the big accounts, I found most had free or low cost entry level services specifically designed for cash strapped emerging businesses like ours. I spoke to vendors from Xenon, Mathworks, Qlik, Snaplogic, Angoss, SAS, Yellowfin, Intech, CBIG, Corsera, CM Systems and Teradata to name a few. There are heaps of free open source solutions on the market as speaker Peter Inge of Caltex alluded to. However, like everything else, data analytics tools are becoming more accessible and easier to use.

The vendors were typically leveraging co-working spaces and accelerators as market channels for start-ups. Start-ups can be great advocates in promoting services, can help with feedback on product development, and you might hit on one that survives more than a year or two. We get a lot of interns working for us that have experience with tools like Matlab coming from renewable energy engineering schools, so that is a great resource to tap into as well.

At the CDAO round table luncheon on day two, I sat next to Mark Hunter CDO of Sainsbury’s Bank, Craig Rogers of SAS, Sue Fong Kong Pricing Lead at Caltex, Carl Teves Analytics Business Process Champion at Lion, and Vitus Chu Enterprise Architect (Information) at RMS. I was the token start-up CDO and had quite a different perspective. However, I have previously spent close to ten years working for First Data, a large multi-national payments processor, running analytics as an internal support function to the business. My take away from the conference is that there are several key concerns that senior data professionals keep coming back to.

 

Takeaways from Corinium's Chief Data and Analytics Officer (CDAO) Sydney Conference

 

1. How to Monetize Data Assets

 

How to monetize large repositories of data was voted as the topic for discussion at the open forum campaign drinks led by Glen Rabie, CEO of Yellowfin. Kate Carruthers CDO of UNSW raised the point that Westfields spend a lot of time looking at their data but couldn’t find anyone to buy it. This is a fairly common story.

During my time at First Data, we looked at various ways data could be monetized. Generally, internally focused projects were successful, and external were not. For example, one of the projects I headed up was to improve cash servicing efficiency of ATMs. In terms of the ‘Internet Of Things’, cash dispensing ATMs track and monitor pretty much everything. The project achieved a ##% operating cost savings in the first year by optimising frequency of cash delivery by armoured vehicles and the amount of cash delivered. At the time, this was credited by the CEO as transformational in the industry.

However, monetizing data through an external facing service was more challenging, particularly considering data privacy and other regulatory constraints. One example comes to mind where we developed a customer facing dashboard for merchant services customers with loyalty analytics and peer benchmarking. However, larger merchants such as supermarket chains already had better analytics. On the other hand smaller merchants like coffee shops or dental practices tend to have a more intimate grasp of their customer base and were not all that interested in payment analytics.

 

2. The Start-up Experience

 

OK … this one wasn’t really a major theme at the conference, but perhaps it could be in the future. Somehow, working in a small start-up with limited resources, we have succeeded in delivering a customer facing data monetization strategy with an award winning product. This perhaps ties into a core message from speaker Matt Kuperholz Chief Data Scientist and Partner at PwC, that you should start with a business problem, not with a technical challenge.

Winner of Strata Community Australia’s “Innovation of the Year” award in 2016, Wattblock has been successful because it solves a big problem. Residential strata buildings are decades behind in energy efficiency and strata committees are largely oblivious to spending up to 4x too much on common area utility bills. City governments are acutely aware of the problem but struggle to get engagement.

The data on Australia’s 400,000 plus residential apartment buildings is sitting in the filing cabinets of some 2,000 strata management offices and in online utility portals. However, most lack the ability to analyse that data efficiently. Wattblock delivers ‘executive summary’ style dashboard reports and insights which strata managers sell to their customer base. This enables strata managers to deliver more value and mobilise more downstream work for themselves.

At the core of Wattblock’s service we have built what I call the ‘Data Hierarchy Model’. That sounds commonplace, but essentially it enables us to benchmark energy usage with as little as two inputs. We can therefore start providing value with minimal customer friction, something we learned from Telstra’s start-up accelerator, muru-D.

With minimal inputs Wattblock is immediately predicting such things as whether or not the building has an indoor heated swimming pool, how much roof area there is for solar panels, the number of residents with electric vehicles, and the number of lights in the corridors, fire escapes, and basement carparking. The data hierarchy works in such a way to target the next most accessible inputs having the most predictive power over the end result.

The service is also provided on subscription which enables the data for each building to be enriched over time. Keeping in mind we have only been running for a few years, for our longest standing customers we not only know the light count, but the precise location, make, model, and settings of every single light, ventilation fan, AC unit, pool filter, hot water tank and lift with real time utility data feeds. So before investing in connecting up the building with smart sensors, we are leveraging basic data that already exists to make buildings smarter.

 

3. Data Visualisation

 

There was an interesting workshop run by Christian Bowman of Ladbrokes on data visualisation at the conference, ironically the only one without any AV support. Good data visualisation practices are a critical component of our reporting service. There is an art to creating clear communication supported by data. Data professionals tend to come from technology, science and statistics backgrounds and have a tendency to produce thorough looking graphs with everything labelled accurately including full descriptions and proper units of measurement. A bit of graphic design sensibility can help to swallow the pill. We take a reductionist approach to data visualisation that focuses on delivering key messages clearly and with minimal effort being placed on the reader. We don’t want the reader sitting there studying a chart trying to find some insight from it. We’ve already done the work to find the key insights, so they should jump off the page.

 

4. Big Data, Machine Learning and Artificial Intelligence

 

These were some of the more exciting buzz words at the conference; artificial intelligence, real time architecting, self driving cars, voice assistants and home automation. In particular I really enjoyed a vendor presentation by Xenon’s CTO and Head of R&D Werner Scholz and Solution Architect Peter McGonigal discussing Nvidia, Mineset, ‘visual data mining’ and machine learning.

This is probably a step removed from the immediate concerns of most Australian corporates. However, I wouldn’t be surprised if the larger banks, utilities and supermarkets have already bought themselves a 250 node Nvidia DGX-1 A.I. supercomputer or two to play with. Amazon has shown the way with their Amazon Go walk-in walk-out retail experience. There are a number of device companies targeting the connected smart city opportunity. However, probably the most interesting thing I have heard about from an Australian company is Resmed’s $800 million acquisition of Brighttree which has the potential to provide doctors with real time sleep diagnostics of their patients. However, apparently this too faces difficult regulatory barriers.

From this perspective I’m inclined to follow and invest in the larger players. Nvidia was my most successful pick last year, and I’m inclined to back it again this year. However I’m also watching Amazon, Google, Microsoft and Apple particularly around voice assistant services. I didn’t make it to the CES show this year, but apparently integration of voice assistants was a major theme. I think we are going to see some big investments starting to play out this year.

 

5. Fostering Innovative Culture

 

Can innovation be driven by a broader embedding of data in the workplace culture? Insights come from observation of the data, which might happen anywhere at any time. This can lead to value creating innovation, if only the right culture were in place to capture it.  There was talk of building cross functional ‘tribes’ and using agile methodologies. However, it was acknowledged that without a constant push people tend to fall back into their role boundaries.

This is not the case in the start-up world. Working in a start-up is do or die week to week and everyone has to pitch in across all facets of the business. On the other hand you also have the flexibility not to do the things that don’t matter. While corporates often have the best talent, start-ups can still foster a culture of high performance none the less.

At the luncheon roundtable I posed the question as to whether or not others thought there was an emerging market for outsourcing ‘disruption’. I’m aware of a few small teams that are peddling start-up culture for hire. There was general agreement that there is a market for it, but mainly on a temporary contract basis. The view was expressed that disruptive culture can still be achieved in time and that outsourcing would only be a short term fix.

Bringing external disruptors into a corporate environment doesn’t always work the way you might hope. The major consideration is that employees may become defensive and feel disenfranchised, overlooked or even cheated. People understand the value they have to offer. Just because you pay someone a wage doesn’t mean they will go the extra mile without any resistance or political pressure. However, where new analytics tools and resources can be explored an outsourced team often makes sense. In other words, creating a sandbox with some new tools and observing what comes out of it. The other popular way to outsource ‘disruption’ is the off-site Design Thinking workshop where contractors are really just event facilitators.

Thanks to Anthony Tockar Decision Scientist at CBIG consulting for hosting a breakfast roundtable on day three. We further discussed outsourcing data analytics services and the challenges large corporates face in making data available for consultants to work with. I asked how they keep the data segregated and maintain ‘Chinese Walls’ and whether or not they have looked at bringing competitive parties together for industry benchmarking roundtables. Apparently this is fairly common in the US and starting to take hold here. Daniel Collins Manager Demand Forecasting and Insights at Origin Energy raised the concern that any outsourced analytics would face multiple barriers just to get access to any data. I also had a chat to Daniel about a customer of ours that has just been handed a $300,000 increase in their annual electricity bill. It seems the energy market is going through some big shocks at the moment. Noted Benny Coscia of the Australian Energy Regulator was also in attendance. Now that is an industry that could benefit from some industry round table discussions right now.

Well those were my top five themes in no particular order. I didn’t get to see everything. Great conference overall and I’m hoping to see it coming back bigger and better next year.

 

By Ross McIntyre, CDO, Wattblock

Save

Save

Save

Save

Save

Related posts