14 Jan Why devops could provide the solution to your data manageme…
Few organisations have been able to match explosive growth in data volumes with their capacity to turn that data into something useful.
With traditional tools and processes struggling to keep pace, many are applying a practice from the software development world to accelerate getting data into the hands of those people who need it. That practice is known as dataops, and it borrows heavily from the devops concept now commonly used to accelerate and streamline software delivery.
While the concept might be new to many organisations, for Simon Raik-Allen, setting up a dataops capability was one of the last jobs he undertook before leaving his role as chief technical advisor at MYOB back in 2017.
“Organisations are producing huge amounts of data, and most of it is falling on the cutting room floor,” Raik-Allen tells CMO. “If you really do want to do something with it then you’ve got to collect it and put it into places where it can be found and used.
“There is so much being collected now that you need a team that actually knows what they are doing with it. It is no longer enough to have a couple of SQL hackers who put it in a database.”
Dataops borrows from the devops by using software development resources (the data/dev side) to make data available to data scientists and others who need it, accompanied by a dedicated approach to the physical machines and networking that host it (the ops component). Raik-Allen says this approach alleviates the problems often experienced when organisations create a centralised data analytics function, which can lose touch with the needs of the organisation.
“The dataops team don’t do any of the analysis,” he says. “What they do is get all the data and make it all available to others in the organisation. And now it is very easy for you to give it to the data analysts and scientists who are being employed in the various division to do their analysis.”
CMO and CIO partnership
It is a concept being embraced in the global operations of data management technology company, Hitachi Vantara, where a partnership between the CMO and CIO to implement dataops has delivered rapid revenue uplift from its targeted sales activity.
Hitachi Vantara CMO, Jonathan Martin, says dataops is already one of the key ideas his company takes to customers, so it made sense for to undertake some self-experimentation. So when CIO, Renee Lahti, was looking for opportunities to apply the company’s AI and analytics technologies and dataops methodology to its own data, Martin was the first to raise his hand.
“With this project, we sought to use data to better predict customer needs so our sales team can come to them with the solutions they need when they need it most – sometimes even before the realise they need it,” Martin says.
Martin and Lahti used the company’s own Digital Value Enablement (DVE) methodology to review the ways it could hone in on target customers, from adapting the marketing mix to targeted selling, conversion rate analysis, cross-sell and upsell, and training and executive briefing.
“Through the exercise, we settled on a specific area that could benefit from dataops – cross-sell and upsell,” Martin says. “This allowed us to drastically shorten the time to value of the project. We identified four key datasets, instead of dozens, which included transaction data, intent data, firmographics data and IoT and usage patterns.
“Through this process, we ultimately identified US$28 million worth of new potential revenue within our customer base.”
While the project yielded a number of insights into the application of dataops generally, Martin says it has also created in a significant cultural change.
“We’ve been able to build a deep partnership between the marketing and IT teams as we worked together on the same goal,” he says. “Our different skillsets and points of view came together beautifully for drastically better results. We were able to move mountains and get a ton of work accomplished in only four or five months.”
Lahti agrees while dataops is a methodology and has a technical aspect, it is also a mindset that focuses the organisation on improving the data use and reducing project cycle times. And it can continue to add more and more value over time.
“Data doesn’t get stale,” Lahti says. “The solutions will change over time, but the data is the data. The thing for me that’s so exciting is that as an executive if you have a data set and bring some algorithms, we can reuse that data. So the cost of finding value continually reduces. That is your dataops advantage.”
Lahti says the dataops approach also alleviates the frustration some organisations experience when pulling their data resources together by enabling the data to remain where it is.
“You spend 70 per cent of your time prepping data to move it to the data lake and less than 25 per cent looking at the data,” she says.
Another critical factor in the success of the project has been the relationship formed between the marketing and IT functions – something she says can be replicated in any organisation.
“Data ops is a mandate by which we can tear down those barriers between us versus them,” she says.
Martin has already developed a series of ideas where he sees dataops as adding value, such as improving Hitachi Vantara’s marketing mix, targeted selling, conversion rate analysis and its cross-sell and up-sell activity.
“As a CMO, I am always looking for ways to better understand our customers and then get the insights needed to best serve their needs,” Martin says. “And with my background in computer science, I am a prototypical data-driven marketer. The idea of applying dataops to better predict customer needs so our sales team can come to them with the solutions they need when they need it most was music to my ears.
“I was on-board from day one and haven’t looked back since.”