Computers and artificial intelligence have come along at an exponential rate over the past few decades, from being regarded as oversized adding machines to the point where they have played integral roles in some legitimately creative endeavours.
Organisations looking to employ marketing analytics should secure stakeholder buy-in upfront and strive for a standardised approach to Return on Marketing Investment (ROMI) if they wish to be truly effective.
That’s the view of Telstra general manager of products and marketing analytics, Rowan Drummond, who shared his unique insights into how the telco giant is employing analytics at the SAS Impact Seminar Series in Sydney on 21 November.
Drummond claimed the telco giant’s own analytics journey over the past two years has taken it from a “brand and ad tracking” business, to utilising data and analytics to optimise future marketing activities and improve customer advocacy. To do this, Telstra’s marketing analytics team has devised what it calls the four new ‘Ps’ of analytics: Precision, process, people and presentation.
“It’s important to get these things right if you want to go from having a metric that smart analyst have developed, to a metric that the business can action,” he told attendees. “There has to be the will to action that.”
The first P, precision, is about having the right and accurate data, along with buy-in from the business around using that data, Drummond explained. The way Telstra’s analytics team has ensured this is by getting teams to sign-off on accuracy of information derived from various sources before it starts providing insights back into the business units.
“A lot of the things you might do will be quite new to the organisation. We present first on how accurate the model was before we present the analytics. We also use data to look back first, to show how accurate we might be moving forward,” he said.
The second step, process, is about informing the business process, Drummond said. “We have tried to roll out a ROMI metric before and the modelling was pretty good, but overall it failed because we didn’t have the buy-in and weren’t able to embed it into the business process,” he said.
“We already have processes in place for allocating marketing budgets, and for reassigning campaign spend. What we’ve done is layered on an extra piece of information. It’s important to make that point clearly upfront. You will get better buy-in, plus less pushback as well.”
People is about having the right team in place to provide insights, as well as getting the rest of the business on-side. To do this, Telstra centralised its analytics function in 2011 into one unit under the office of the CMO, Mark Buckman. The team includes analytics, customer modelling, research, competitive intelligence and data and technology team underpinning it.
“We own this process and the metric that comes out as part of the CMO function,” Drummond said. “Because we are in the CMO, it’s also easier to engage with our partners – all those marketing managers that we need to get across the line. You need to get them on-board early on.”
Drummond’s fourth P, presentation, is about translating complex analytics processes into a story that’s easily digestible for business stakeholders, Drummond said. To do this, Telstra has rolled out a dashboard tool, based on Accenture’s Agile Marketing Analytics Platform (AMAP), to marketing managers as well as media buying partners so they can run their marketing analysis and optimisations on activities the fly. All analytics is fed into the tool, which also features the current marketing plans.
“A lot of our marketing is quite cyclical… if we can look at how that went in a particular area and particular medium channel last year, we can predict quite well what will happen this year,” Drummond said. “The marketers can then use this tool to tweak spend to get the best return.”
While ROMI is a key priority for Telstra’s analytics team, the other area of focus is helping to deliver the customer-centric approach espoused by Telstra’s CMO. This saw the telco adopt the net promoter system (NPS) as the basis for all of its marketing efforts two years ago.
Drummond revealed his team is working on evolving customer segmentation from channel-specific modelling, to an ‘uplift’ modelling approach which will better tackle customer lifetime value and profitability.
Unlike traditional response modelling, uplift modelling is about predicting the behaviour both of a treated group as well as a controlled group of customers to an action. This is based on four groupings – the persuadables (those who only respond when they’re targeted); sure things (customer that respond whether they’re targeted or not); lost causes (those who won’t respond irrespective of your activity); and sleeping dogs (customers less likely to respond to a direct marketing action).
One area Telstra is adopting this approach is around customer churn. Drummond said it is also looking to roll this out across all of its customer modelling.
“This model helps us to split out the sure things from the persuadables, which in traditional response modelling is easy to confuse,” he said. “It means you end up contacting customers that you do need to contact, and we can get a lot more targeted with what we’re doing.
“It also makes it clearer what customer we shouldn’t be targeting.”