8 things we learnt about big data analytics from the Adobe Summit
- 22 April, 2014 12:12
Big data is big news, and nowhere was this more apparent than at this year’s Adobe Digital Marketing Summit in Salt Lake City.
Here, we list eight things we took away about big data and data analytics from the event.
1. Organisations don’t agree on the best way to deal with data
Data might be the common problem all organisations face in their quest to be more competitive, customer-centric, efficient and innovative, but the approach to tackling it is widely diverse.
As was made plain in a panel dispute between Fox Sports and Turner Broadcasting during Adobe Summit 2014, there is no clear agreement on how organisations handle digital disruption and the data explosion that has come with it. Representatives from the two sides sparred over whether to adopt a ‘hub and spoke’ approach to data analytics by investing in centralised, specialist teams, or to bring data insights and analytics capability into existing teams to better utilise.
There was also debate on when to invest in technology, what an ‘operating system of data’ actually entails, innovation versus disrupting existing revenue streams, and the extent to which organisations are able to rely on data insights.
“Traditionally, the people looking at the churn data sit in the research teams, in brands, ad serving teams. We recognised that to understand the physics and how we were engaging with customers, we had to bring that information together,” explained turner Broadcasting Systems’ senior director of analytics product strategy and data governance, Colin Coleman.
“You can’t start integrating data together unless you’re standardised. The unsexy part of this is building that foundation on top for a house of data. That was a big upfront investment, but necessary to unlock how to bring all that data together.”
But until organisations have the “operating system of data”, and see the clear correlation between data insights and significant bottom-line impact, Fox Sports chief digital officer, Jeff Misenti, said data utilisation is about serving the audience as best it can and protecting revenue streams.
“As much we believe the data will guide us into doing things, we are still making great leaps of faith,” he said. “We’re collecting all this information in the hopes it will lead us to new things, but so much is still based on gut feel.
“We haven’t found the [data] haystack, so how can you expect to find the needle?”
2. There is an overarching framework you can utilise to make data actionable
Of course, a difference of opinion doesn’t mean that there aren’t clear steps you can take to improve data utilisation and, more importantly, take action.
According to Adobe, the five steps marketers should be using to improve their chances of actually getting value out of data for business decision-making are:
- Ingest – Connect, store and merge data in an organised fashion
- Distill insights– Turn that data into insights and put it through processes so you can do something with it
- Curate - Condense data insights into consumable packages for the business to use
- Syndicate - Deliver data back to the lines of business
- Optimise - Personalise data for different audiences, test and learn
Related: 8 ways to get on top of data analytics
3. General knowledge about big data is diverse and often contradictory
Big data means many different things to many different people. While the official definition of big data is the four ‘Vs’ – volume, variety, velocity and veracity – it’s clear many marketers and companies are using the term to describe everything from basic-level analytics and reporting on internal data sets, through to real-time responsiveness using third-party data consumption and predictive analytics.
For Live Nation’s director of strategy and analytics, Jonathan Watts, who spoke on the Thinking Small: A panel on making big data actionable panel, big data ultimately comes back to the business challenges he faces. “I don’t really use the term big data; for me it describes a problem I run into,” he said.
“Usually when I’m dealing with a specific data set, I can generally work with that within the tool that I have. But when you start to try and cobble together data sets from different sources, you run into a problem where the tools don’t work anymore. That’s what big data is to me – when you start to try and combine things and you run into processing and technology power issues that you need to solve it.”
Big data can be all different sizes depending on your business, said Digital Clarity Group president and principal analyst, Scott Liewehr. He described big data today as the “phenomenon of utilising data to make more informed business decisions”.
“Up until a year ago, I had been using the official definition of big data – which is that it’s data so large your databases can’t handle it,” he said. “But I’ve stopped focusing on the true definition of the term and made it akin to ‘mobile first’ – it doesn’t mean you have to have a mobile strategy first, but it started to make us think from the perspective of the customer.
“If we are thinking about using data, gathering insights about our customers, processes and products and making informed decisions based on that data, that’s a good thing.”
4. Data is only as useful as your ability to action it
Whatever big data means to you, it’s useless without the business case, processes and people in place to make it actionable.
In a customer video during the Thinking Small: A panel on making big data actionable panel session, Apollo Group’s June Dershewitz highlighted the difference between the buzz of big data, and the reality it should be presenting for businesses. She used an example of hourly alerts on her company’s website as something that helps maintain revenue and be more responsive to customers.
“We need the data to ask the questions about what is important to us as a business, and how we can then use data to help achieve those objectives,” Dershewitz said. “There’s a lot of conversation about real-time, but you only actually need data as quickly as you can turn around and take action on it. But if we can act on the data quickly, we should be able to get access to the data quickly.”
One of the challenges many organisations have in common is where to start with big data. Digital Clarity Group’s Scott Liewehr suggested the best place to get started is by utilising data insights to optimise what you’re doing already.
“How can you support what you’re doing already and optimise that, as well as do it better,” he asked. “For example, if you’ve started a process to segment the audience in some way: What data can help you understand whether or not the segments are in fact accurate? How can I take what I’m doing, where I’m making some assumptions, and validate those through data? That’s a relatively easy first step to bringing data-driven decision making into your organisation.”
Live Nation’s Jonathan Watts also outlined a three-step process his company uses to make sure it’s actioning data well:
- Decide what’s important: Choose what you can test. “What are the big questions that can drive your company forward,” he advised. “Don’t even think about the data; just ask what are the things that can make a difference to your business, then worry about the data limitations later.”
- Create and test smaller hypotheses using ‘medium-sized’ data tests: This will give you a good idea of whether the answer is A or B. “It doesn’t have to be perfect,” Watts adds. “Use available data and don’t wait”.
- Build a production ‘big data solution’. This tool set and insights should then be democratised through your organisation to be used most effectively, Watts said.
To ensure her organisation has the people and process in place to utilise data insights, AOL’s senior marketing director, Jennifer Towns, said the group has brought testing and optimisation into every aspect of consumer engagement and content delivery.
“What we know about our customers is driving the experiences, offers, products as are giving them, and the content. And it’s very personalised,” Towns said. “When we do this, we see huge increases in our core KPI. And at any point in time, if you deliver a lift on an experience, reduce churn, or drive more search volume, it’s a very easy sell to the executives that data matters.”
Page Break
5. Technology investment is a must if you’re hoping to utilise data successfully
To really get on top of multiple data sets, however, you are going to need technology; and lots of it. Lenovo director of global business intelligence, Ashish Braganza, outlined how the PC manufacturer has created a big data technology stack and customer intelligence platform nicknamed ‘Luci in the Sky’ (Lenovo Unified Customer Intelligence platform) to tackle the challenges of processing massive swathes of customer data from more than 60 data sources daily into actionable insight. These data sets include SAP, the Lenovo.com website, customer surveys, call centre information and various other sources.
Based on the Hadoop file system, the list of technologies within Lenovo’s stack is considerable and stretches from a data integration platform layer, through to the LUCI cloud and internal platform. On top of this are a range of visualisation and predictive analytics tools including Tableau Software, IBM SPSS, SAS Enterprise Miner and Adobe Insight.
The businesses supported by Lenovo’s big data stack include online customer behaviour, social media analytics, customer segmentation, profitability, campaign performance and marketing dashboards.
“We didn’t get to have a big data technology platform at once; we went through a maturity curve at Lenovo,” Braganza pointed out. “A lot of organisations start with Excel, and MS Access. Pre-2007, we were doing a lot of SQL… then in fiscal year 2012, we were on a no SQL database structure. At the beginning of this year, we launched our LUCI platform.”
6. Data-driven marketing does work
Hard examples of how data is being used effectively by brands are still few and far between, largely because those who have worked out how to make data an asset don’t want to give their competitive advantage away.
From what we could gather at several case study sessions, though, being able to process information and use it to modernise customer engagement does work.
As just one example, US-based beauty retailer, Ulta Beauty, shared how it is using data to fuel campaign management and drive stronger customer loyalty. Historically, the marketing team had to ask IT for information from 38 siloed customer databases across seven individual systems, making responsive marketing a challenge.
Unifying these databases and giving it a single view of loyalty program holders has helped to drive up EDM open rates, and increase sales per delivery by 141 per cent.
7. Data-driven marketing is transforming the role and responsibilities of the CMO
Adobe CMO, Ann Lewnes, pointed to the data-driven nature of marketing as the catalyst for the rise and transformation of the chief marketing officer.
“One of my favourite quotes from our CEO, Shantanu Narayen, is that he expects the CMO to soon have a better pulse on the business than the CFO because we’re so immersed in the numbers,” she commented.
For Audi of America general manager of digital technology and strategy, Jeff Titus, embracing agile ways of working and quicker product cycles are vital for a modern CMO. He also stressed data know-how as one of the most important tools in the marketer’s arsenal today.
“Every marketer needs to understand that when you put something out there, you should have to measure it,” he said. “That’s not referring back to a chart, but in as real-time a way as possible. Get the team oriented to that, and look at outcomes, not just the journey and what might happen and don’t try to segment down to the last level.”
8. Big data is like teenage sex
Again and again, speakers during the Adobe Summit used Dan Arely’s quote comparing big data to teenage sex to illustrate the state of knowledge today: 'Everyone talks about it, nobody really knows how to do it, everyone thinks everyone else is doing it, so everyone claims they are doing it.'
What should marketers read into that? CMO’s view is that most organisations recognise there is still a lot to learn before they truly get a handle on big data – even though most of them won’t admit it.
So if you are still struggling with how to best capitalise data-driven decision marketing in your organisation take heart; you’re not alone. But what’s equally clear is that those who aren’t taking those first steps now, could well end up 30-year old virgins.
Follow CMO on Twitter: @CMOAustralia, take part in the CMO Australia conversation on LinkedIn: CMO Australia, or join us on Facebook: https://www.facebook.com/CMOAustralia
