Mobile ads in 2016: Big data meets the big idea

Aurelie Guerrieri is vice-president of global marketing solutions at Cheetah Ad Platform

Mobile Advertising meets Data Analytics

Mobile advertising has reached a point where big data is beginning to line up with the big idea: native ads and content exquisitely targeted to the mobile user’s data signals, screen and circumstances.

The key to success is a focus on in-app behavioral signals to determine if the user is a candidate to make a mobile purchase, and of what type, and when. But sorting out which of the mobile native’s millions of data signals have the potential to yield the greatest return, then using technology to action that intelligence at scale, is tricky.

The haystacks are large, the needles are few and unique

At the time of writing, there are about 2 billion mobile internet users on the planet. As you can imagine, all that activity produces an enormous amount of data, a lot of it from inside of apps where data signals contain behavioural information of tremendous value to advertisers. The problem is that mobile users generate so much data that it is impossible for anyone to make sense of it all.

For advertisers to get down to what matters in engaging the mobile native, they first need to adopt a less is more approach and commit themselves to concentrating only on the signals that are indicative of what moves me down the path towards a transaction. Let me emphasize the word ‘me’ in the last sentence, because even if 10,000 other women around the world are interested in buying the same Burberry Brit trench coat as me, if you want my business, you’re going to have to find a message and a moment that are just right to get me to buy, not just repurpose a generic ad and drop it on my screen whenever you feel like it.

It takes good quality data to get good results

First-party data is essential – there is zero point attempting to create targeted native content from third-party data with no behavioural value that relies on ‘bucket think’ approaches to segmentation.

It’s also imperative that the data advertisers use is fresh. You don’t want to put someone in a category as a ‘great potential buyer’ based on data that’s old and may not really be indicative of anything about the me of today. So what if I downloaded Candy Crush six months ago? Maybe I downloaded it to entertain one of my kids on a long car trip. Don’t make a mistake of painting me with the label of ‘casual gamer’ just because that fits some easy marketing definition.

Algorithms tend to come in a black box

As I addressed above, marketers have little ability to parse how good somebody else’s data is. In the same vein, it is virtually impossible to know how good another person’s algorithms for targeting or latching are.

A quick analysis of my personal data would reveal I like to play online poker games. Typically, that’s an activity that marketers associate with men. So, if you were using an algorithm to select targets for your new poker app, the technology would probably assume that I’m a man and serve me up an ad with a buxom girl, or colour and font combinations that I’d find garish, to try and get me to download, and that will fail miserably.

There are data verification tools and strategies out there to help brands validate third party data and algorithms, but those solutions are nascent and unproven.

Moments matter, context is king

In order to meet each of these challenges, it’s important for brands that want to succeed in mobile—and in m-commerce in particular—to develop the skills to leverage in-app data. Because, even more importantly than raw behavioural data, in-app data provides the advertiser with the context required to understand my total experience as a mobile native.

I did play a poker app for 25 minutes last night before bedtime. I did open and close the app twice while responding to messages from friends on Facebook, and I did click on an ad for the elusive trench coat that I’m looking for while I was on Facebook. If the advertiser isn’t working from a position of having my fresh, in-app mobile data signals in hand, how would they ever know any of that?

Amazon Prime has the right idea

Recently, Amazon introduced Prime Now, an app that combines shopping, entertainment and everything else all-in-one. Anything that you ever see on Amazon—on desktop, mobile web or in-app—is accessible from Prime Now.

It’s nice that I don’t need to go back to my desktop or another app to shop, but it’s even nicer that I don’t need to make a trip to the store to buy groceries. I don’t even need a shopping list. I just open my Prime Now, drop things in my shopping cart and it’s at my house an hour later. This is a real-life use case of a brand with a heavy stake in the future of m-commerce using a combination of data signals, algorithms and slick native calls to action to make my life better and earn a bit of my money in the process.


The big idea in mobile native advertising boils down to using small parts of user’s mobile data signals to provide calls to action that are delivered at just the right moment, in just the right context, to inspire an m-commerce transaction.

Where before this was just an idea or wishful  thinking, the confluence of trillions of first party in-app data signals, the technology to interpret what they mean, and the commitment to treat each mobile native like the unique individual we are is starting to make advertiser’s big ideas for big data a reality.

Experience Digital by Mike Kent

Marketers should be ‘customer experience designers’

The role of a marketer should be redubbed a “customer experience designer”, according to Merlin Entertainment’s group marketing director, Emma Woods, who said the ability to deliver good consumer experiences is even harder in a tech-driven world where things can easily go wrong.

Woods, speaking at Experian’s client summit in London today (30th September), said that the window for delivering customer experiences is narrowing and the correct use of data is ever more important.

“The world of marketing is changing and we have to think about ourselves as customer experience designers,” she said. “Part of that design responsibility is also being guardians for when things go wrong. The second thing is that… in the future your customer data will inform that customer experience and my challenge to marketers is that you have a responsibility to collect it, use it and nurture it.”

A recent step to improve the various customer touch points at Merlin Entertainment – owner of Legoland, Alton Towers and Madame Tussauds – is the release of a new Legoland app which directs visitors to attractions with fewer queues, the nearest restaurant and delivers real-time information about queuing time.

Since its roll out in July Woods, who was brought on board in 2013 to broaden the digital journey of the company’s 55 million guests, said that 15 per cent of visitors have downloaded the app and reported a less stressful experience, particularly on extremely busy days.

“We need to be meticulous about understanding all the customer touch points and thinking about what is the experience that the customer wants and how can we facilitate that through great service or technology?” She added.

Also speaking was Jon Wilkins, executive chairman at Karmarama, who lamented the advent of data companies which he said have caused a “problematic” relationship with creativity, which he likened to a “straight jacket”. He used the example of Netflix-created House of Cards where at a data conference the streaming site’s chief executive Reed Hastings told producer David Fincher that he should consider a data insight for future shows that showed a certain point where viewers switched off.

“His response was ‘never tell me that again’ and that’s a standard discussion between data and creativity,” said Wilkins.

To ease the friction creativity should be “tech driven rather than tech led”, an idea that connects with the role of data and how it can inspire creativity.

Experience Digital by Mike Kent

There is a real need to develop a solution for Personalised Discovery


Personalised Discovery - Can it truly exist?

What do you do when you don’t know what you want to read, watch, listen to or do next? What do you do if you don’t know what to search for? Or can’t describe clearly what you’d be interested in next?

There are so many great choices available in the digital realm, and new stuff is pouring in every second. Many times we feel helpless in front of such an abundance of endless possibilities.

Nevertheless, so far no one has created a solution that would automatically bring all the interesting options to your fingertips without you asking for it. A universal personalized Discovery solution doesn’t exist yet. Why?

Personalized Discovery Today

There have been various attempts and approaches to crack personalized Discovery — at least partially.

StumbleUpon has been around for a while. The app provides content based on selected categories and other “Stumblers” you follow. Flipboard’s personalized magazine has transformed into a social news platform. You personalize your own experience by curating content sources and following people. Pinterest, too, has a follow model for people, their content and topics. Its Guided Search with combinable keywords works as an additional interface alongside the curated feed.

Pocket recently released its Recommended section that provides content based on the things that you saved for later. Google Now delivers useful information based on your previous actions and historical data. And Facebook is just entering the game with its M that supposedly recommends actions and content.

Finding the right dynamics for personalized Discovery could be the key for creating more human-centered and diverse digital experiences for all of us.

Most of today’s Discovery solutions resemble social media’s friendship or follow paradigm. You follow people and their content or you follow selected keywords and categories to personalize your own experience. However, this paradigm tends to reinforce our existing information silos. By directly customizing things for ourselves, our social and personal biases restrict the way we expose ourselves to new information.

Additionally, it becomes hard to estimate how much personalization actually happens automatically and how it helps us in discovering things. The wider the pool of information, the more we have to work to detect the signal from the noise.

At the same time, the user experience of a purely machine-powered approach hasn’t still crossed the “uncanny valley.” A machine telling us what we should be seeing and doing next has a dystopian aura, even in very mundane circumstances. Many times the algorithmic suggestions that don’t directly reflect our social environment or past interactions appear to be too obtrusive or outright ridiculous. Indeed, machine-powered Google Now focuses on delivering useful information instead of new explorative choices.

Discovery currently exists as a category of Internet services and apps, as well as a functional feature in some content-specific platforms (e.g., Spotify and Snapchat). But no one has come up with a universal Discovery paradigm for the mainstream audience.

The Unsolved Discovery Puzzle

There are five main challenges faced by the universal Discovery solution:

Clear Value Proposition And Use Case. Universal Discovery lacks a clearly defined value proposition, and thus a crystal clear use case. Why do you need universal Discovery in the first place? In Search we are looking for specific and relevant answers. In Social we’re connecting and communicating with other people. But how do you define a universal value proposition and use case for something so highly subjective and contextual as Discovery?

People don’t think consciously that they are “discovering” something. They might not even recognize a need for “discovery.” On the contrary, a discovery happens often as a by-product of some other activity.

Frictionless User Experience. Current mainstream user interfaces and experiences haven’t been designed, developed or optimized for Discovery. For example, the news feed and its variations provide a very linear and limited way of presenting information. Personalized Discovery requires a new design approach because finding interesting choices includes potential effort and friction.

Trial and error can form a significant part of the exploration process. Friction emerges when we encounter unexpected choices or when we need to wait for something to happen. Additionally, the experience should proactively pique our curiosity, simultaneously outweighing our personal and social biases.

Technologies For Adaptive Personalization And Content Presentation. Creating a universal Discovery solution brings together two major technological challenges: unobtrusive personalization and sleek content delivery. Adaptive personalization requires an unseen level of automated customization based on our intricate selves. To achieve this, the system needs to be able to capture our meaningful interactions and utilize our diverse personal data.

Current applications of personalization using human curation, algorithmic systems and machine learning methods — or their combinations — don’t yet learn or deliver fast enough, nor do they let us express ourselves as unique individuals. Concurrently, the various forms and types of digital content are messy, and require a lot of sophisticated processing to be presented fluently in various screens and devices.

Accessible Data And Content. The almost infinite sea we call the Internet has become a collection of confined ponds with their own walls and rules. Platforms build their own understanding of you, and usually they don’t let you control how your data could be used for your own benefit in other places.

Simultaneously, an increasing amount of content is becoming platform-exclusive. Major social platforms are becoming content silos, enabling exploration on their own terms and only inside their own boundaries. Media houses are locking their content behind specific access points.

Our social and personal biases restrict the way we expose ourselves to new information.

Discovery Paradox. Additionally, there’s an inherent tension in combining personalization and Discovery. Personalization is about customizing your experience, guiding your choices and serving information based on your needs and personal preferences. Then again, discovery refers to the things that are somehow new and surprising. Indeed, discovery can be as much about questions as it is about answers. It can be as much about irrational and serendipitous as it is about rational and relevant.

The things that you recognize as meaningful discoveries aren’t necessarily what you expect them to be. You might encounter something you didn’t know you wanted or you didn’t even know existed. A discovery can be very personal and context-specific, thus being meaningful only to you.

So, is there a way to overcome this multitude of complex challenges? Or is the universal Discovery solution just a Fata Morgana of the early age of personalization?

How Could Personalized Discovery Work?

A truly smart universal Discovery system makes sense. The amount of information is exploding, and we need better methods to make sense of it. At the same time, the current tools provide only a restricted access to the information that is beyond our personal and social bubbles.

Personalized Discovery can find a balance between relevance and serendipity, as well as rewards and friction, by creating favorable discovery conditions for you as a unique individual. The system understands your articulated and ambient interests by mapping the unique connections you see around you.

A universal Discovery system provides choices instead of the one specific answer. By understanding your interests, the system can expose you to things that you find surprising, even challenging. Simultaneously, it provides meaningful information in easily digestible chunks that let you choose your preferred level of engagement. The presentation is modified based on content form, type and context. To serve content from diverse sources, the Discovery system taps into the free as well as paid content pools on your behalf.

As the amount of potentially discoverable information is almost infinite, human curation and algorithmic methods are used to complement each other. Curation can be made a seamless part of the basic Discovery flow. Your actions curate content for other people and educate the system at the same time. The nuanced human assessment of quality is thus interwoven to the machine-powered dynamics such as prioritizing recommendations and presenting information.

A universal personalized Discovery solution doesn’t exist yet. Why?

In Discovery, goal-driven and casual experiences can coexist. The system brings together various content and action categories such as books, music, movies, travel, food, dating and news. By understanding your preferences with movies and the latest pop culture news — and being able to detect your current mood — the Discovery system recommends new interesting music choices. Also, it can surface interesting weak signals and unseen opportunities. By understanding your daily activities, it can serve a surprising micro-eureka moment when you get trapped in your mundane routines.

Maybe Discovery itself is an ambient system. It’s present and available in the background, only activating when it makes sense to you. Time-consuming complex stuff is hidden under the hood. The system notifies you when something is happening or already waiting for you. Such a Discovery solution is your never-sleeping intelligent extension that doesn’t need continuous actions from your part.

This would be more in tune with our natural experience of discovering new, interesting things almost coincidentally. When digital and physical become more and more entangled, ambient Discovery can be the new user experience paradigm for VR.

However, could any Discovery technology help us to find anything truly new and meaningful if we’re not open to exploration ourselves? Maybe a well-tuned Discovery system could educate us to be more open toward diversity and serendipity. And, potentially, finding the right dynamics for personalized Discovery could be the key for creating more human-centered and diverse digital experiences for all of us.

Experience Digital by Mike Kent

The Case For Hosted Exchange

Though Microsoft Exchange with the Outlook client is the most popular email combination in business, it’s also cost and time prohibitive for many businesses to run internally. The level of complexity, its mission-critical nature and the potential for cost savings make Exchange an attractive application to move to the cloud. Choosing Hosted Exchange, as detailed in the recently updated whitepaper, The Case for Hosted Exchange, dramatically reduces the deployment costs of Exchange and offloads most of the administrative duties, like spam filtering and backups, to a trusted cloud provider. The infographic below takes a quick look at the high-level reasons why Hosted Exchange is the best choice for business email and how to get started.


Rackspace® — The Case For Hosted Exchange [Infographic] Rackspace® — The Case For Hosted Exchange [Infographic]

Experience Digital by Mike Kent
Experience Digital by Mike Kent
Experience Digital by Mike Kent
Experience Digital by Mike Kent
Experience Digital by Mike Kent
Experience Digital by Mike Kent
Experience Digital by Mike Kent