Globys targets mobile CMOs with platform update

Big data analytics solutions provider Globys today updated its Mobile Occasions platform that it said puts “Google-like marketing experimentation and dynamic machine learning into the hands of [chief marketing officers] at operators worldwide.”
The company noted the solution is designed to help marketing teams better understand and act upon “customer behaviors, intentions and motivations” in real-time and using the appropriate communication method.
Globys Chief Scientist Dr. Olly Downs provided a run down on the impact the new offering will have on the marketing of services by mobile operators.
RCR Wireless News: Tell us about the capabilities and major features of this new release.
Dr. Olly Downs: What Globys is bringing to bear with this latest release of Mobile Occasions is the ability for Google-like marketing experimentation and dynamic machine learning to be accessible to [chief marketing officers] at operators worldwide. What we’ve done is integrate three core capabilities: automated behavioral analytics for discovery; adaptive always-on experimentation with one-to-one targeting of context based offers for execution; and dynamic machine learning for optimization – into a closed-loop process. This helps CMOs to overcome the challenge around hiring the expertise and developing the technologies which are capable of driving one-to-one personalized experiences and doing so in a way that is scalable to millions of customers and tens of thousands of marketing campaigns.
RCRWN: Can you give us some context for this concept of data science automation?
Downs: We’re seeing mobile operators make significant advancements in big data tools and technologies, but for many the challenge of ingesting, analyzing, and most importantly, acting on the data, remains a challenge. This is particularly true for today’s CMO who is focused on how to better understand customers and apply data-driven insights to improve the customer experience and maximize overall lifetime value.
We’ve seen that the use of big data analytics and scientific marketing can help a CMO transform how they market to and engage with their customers. And yet while scientific marketing is known for its ability to turn high volumes of transactional data into valuable insights, for most marketing teams the process remains very unwieldy – a slow, intuition-driven approach that involves coming up with audiences, white lists, segmentation schemes, and so on, and then designing an experiment to test a very small cardinality of combinations of things.
At Globys we’re focused on helping mobile operators adopt a more scientific approach to marketing and this involves automating the parts of the equation that are small scale or low complexity to enable faster, smarter, and easier marketing. Following a traditional approach, marketers have to think of attributes or audiences and represent them in a campaign spreadsheet or a relatively simple decision tree to express the campaign activity. What the automation allows you to do is set that part aside, apply the marketing intuition to the creative assets of the marketing campaigns and then start testing and exploring the audiences or sub-audiences for which those creative assets actually work. Based upon uncovering insights around what assets work and when, and for what sub-elements of the audience, you move into a loop that is pure: insights, creative thinking, coming up with new assets to test. We think of it as moving to ‘informed intuition.’
RCRWN: What is the goal of data science automation, and how is it different from data science with a human element?
Downs: What operators are telling us is that they need to be able to test more in order to learn more in order to drive better results systematically over time. The impact of transitioning to an approach that relies on data science and leverages automation is two orders of magnitude in the increase of the number of questions you can test and answer – and a significant advantage when trying to determine what works, what doesn’t work, and why.
The way you get a statistically significant measurement – in testing one thing versus another – is having enough individual trials involving each tested element. The other requirement is being able to measure the isolated impact, wherein lies the problem for most operators. With millions of dynamic customers and hundreds or thousands of different treatment combinations, it becomes an impossible process for humans to manage. In addition, it’s not uncommon for operators to be testing a variety of campaign strategies that overlap in terms of encompassed customers. In this scenario, marketers are challenged to efficiently maintain in-context control groups and make sense of the resulting impacts, and ultimately must contend with limited understanding and potentially ‘contaminated’ results.
One of the things our big data marketing system does is automate the rules and constraints that need to be applied in managing the control groups, which allows for the measurement of isolated experimentation across millions of customers and more importantly, unlocks the ability to measure long-term KPIs. What we’ve found is that the reason marketers are often focused on short-term KPIs, like response or immediate purchase, is because they lack the tools to persist a control group for a long period of time or isolate an audience for a certain period of time to measure success. Just looking at the volume of response, marketers have a primary measure that doesn’t have a lot of the noise associated with it but at the same time, it only has indirect connectivity or may even have negative correlation with the ultimate goal the marketer is trying to achieve. This is a significant gap for operators because if you can’t isolate those audiences over an extended period of time, you can’t actually optimize your marketing activity – even the marketing activity that’s manually driven – toward longer term goals.
RCRWN: How do you think the role of the data scientist will evolve in telecom?
Mobile operators, and specifically CMOs, are beginning to view scientific marketing as the next transformational wave of mobile marketing. For many, the way it works today is that there are some broader questions posed to the data scientist regarding the customer audiences: “can she find these behaviors; or can she find customers in these groups; and then the intuition or hypotheses are tested. I don’t think this has to change – but the rate at which the experimentation occurs will drastically increase. Operators will continue to have their data scientists – either as an internal resource or a partner – create those additional audiences or attributes but with the automation in play, they can do it ongoing at enormous scale. They don’t have to stop and wait for an experiment to move to a certain point. They can just set that as a decisioning attribute into an ongoing flow of campaign activity.
By automating the discovery of insights, as well as the execution and optimization of context-based marketing treatments, new technologies will help data scientists become more focused on sophisticated audience definitions that perhaps are unique to the business or to a certain structure of a rate plan or device that the operator is trying to launch. This ability to create, test, adapt and scale complex scenarios that align directly to the customer base will offer significant advantage for the marketer focused on delivering personalized, one-to-one customer experiences based on customers’ needs.
RCRWN: How does the ability to test many offers quickly, and to scale to tens or hundreds of millions of customers, change the game for mobile marketing?
With a system that can scale the adaptive, always-on experimentation of communications and offers, and combine that with dynamic machine learning capabilities, marketers become much better informed about how their products and their interactions relate to the customer. For mobile operators, marketers gain unprecedented insight into how their customers behave and what they respond to which then fuels more granular, dynamic targeting of context-based offers.
What we’re seeing is that it’s game changing in terms of results. With the ability to apply this level of sophistication our clients in the telecom operator space are realizing dramatic double-digit increases in both customer revenues and retention.
We’re also beginning to see how this scientific test and learn approach is shifting how the CMO is perceived throughout the organization. What’s common today with operators is that the CFO and product organization are often feeding the information – by leveraging slow, BI activity – and marketing is simply a recipient. With the ability to increase the complexity and scale of experiments, the CMO is in a position to tell the finance and product teams what’s causing the financial results – what it is about the product, the customer base, the changes in behavior, etc., making them a thought leader in the area of applied analytics and data-driven marketing.
RCR Wireless News also recently spoke with Globys CEO Derek Edwards about trends and capabilities in big data analytics for telecom service providers.