Thinking out of the idiot-box: Advertising via Cloud

MUMBAI, INDIA: Cloud is making inroads into almost every conceivable part of our lives and if Ashay Padwal, Co-founder & CTO, Vserv, is able to execute his vision soon then advertisers and marketers would be witnessing a new language and strategy to talk to consumers beyond the ‘reach’ mindset. At the same time, pricing, standardization and complexity are constraints that need to be sorted out before the potential translates into that ‘wow’ of unprecedented reality. More on why and how brands need to munch big data in this interview ahead.

Tell us something about your concept and model as well as the business and technology curve here?

Vserv is spearheading the trend of using cloud to manage the big data deluge. Our company processes big data from multiple sources and transforms this information into smart data which include user personas, their previous buying behaviour, interests among several other parameters. All this information that Vserv provides to marketers would result in them receiving better ROI, as they would be able to target the right kind of audience for their products. We cater to app developers, publishers, advertisers and telecom providers and are using an intersected cloud and big data approach to support our business growth.

What is the underlying technology that works in this model?

It would not have been possible to serve 2.4 billion ad requests with 1.6 TB of data generated per day without a cloud infrastructure. We have been using cloud since the inception of Vserv, and have adopted big data as well. As the size of our business grew, we required auto scaling for our existing infrastructure in order to cater to the increase in the amount of data that had to be processed. One such cloud solutions company i.e. Amazon Web Services was able to provide us with the ability to scale our infrastructure, based on our requirements.

How has Amazon strategy worked so far?

Amazon has three cloud models, out of which we primarily use Spot instances. Spot instances allow us to name our own price for Amazon’s cloud infrastructure, based on bidding on spare servers. Another key capability we have built on top of AWS is Auto Scaling. Auto scaling allows us to scale our existing capacity up or down automatically, according to the conditions that we define. We use a mix of reactive and predictive algorithms that allow us to structure our infrastructure on an hourly basis.

What implications would an advertising exchange have for agencies, media planners and brands?

Traditional agencies and brands often look at the mobile medium from a reach and not a big data standpoint. They will not fully benefit from this medium until they understand the power of data and its effective utilisation. The scope of mobile will increase with better understanding of user profiles and ‘personas’ of consumers. The mobile medium will thrive once the focus shifts from advertising to marketing as it offers great potential to engage with consumers and allows brands and agencies to gain a different perspective through its use. Agencies and media planners need to strategize and create the right engagements to generate an appealing brand connect and create a loyal customer bank.

Does this not entail the challenge of filtering big data deluge into relevant stuff?

The large amounts of big data that we collect from various sources, is crunched and summarized through our proprietary algorithms, thereby converting this data into actionable data, which is a vital part of our processing goal. The actionable data that is derived from this conversion process enables us in figuring out three key aspects; the right offer/content, right messaging and right value. Based on these three parameters, we are then able to choose the right ad for the user and price it correctly. Also, for us to better understand this data, it is vital for us to remove noise from it.

How?

Noise basically includes data that is based on certain circumstances that are not valid. For this purpose, we have a team of data scientists that helps us to tackle the challenge of evolving the algorithm parameters.

Any trends that you are catching in current app ecosystem and cloud wars?

The app ecosystem will begin to overlap with the cloud. We are already seeing the emergence of several types of sensors and devices from wearables to appliances that incorporate them. These sensors will rely on a number of apps that will provide the user with actionable information. Simultaneously, this information will be sent to the cloud, forming big data on an individual and collective level. An example of cumulative and specific data would be a person knowing how much he has exercised in relation to the rest of the city or other demographic. Along with sensors, OTT messaging apps will continue to grow and become more detailed in nature. Video and more data intensive elements will become a commonplace. As the flow of data increases, the amount stored on the cloud will increase as well, with apps trying to understand and profile users based on their communication habits.

Does this cover gaming space as well?

On the gaming front, apps will become more dynamic instead of the traditional static game play. Game progress is already being synced across devices through the cloud as well as encouraging multiplayer game play. The future will see further connected apps that aim to reflect dynamic circumstances such as weather, events, player habits etc.

Are there any major forces to take into account here?

One of the issues we face with regard to cloud includes the fact that there is no standardization in cloud. Cloud computing has become an increasingly popular approach in recent years, with seemingly nothing but on-going growth that we see in the foreseeable future. However, this rapid growth is today threatened by, the failure of comprehensive cloud-computing standards to gain traction, despite the many groups working on them. A complete lack of standards could make cloud computing more complex to use. For instance the lack of standardization could keep a customer trying to switch from a private to a public cloud from doing so as seamlessly as switching browsers or e-mail systems. In addition, it would keep users from knowing the basic capabilities they could expect from any cloud service. The second issue confronting the cloud industry is that there is no standardization of the cloud API’s (Application programming interface). If there was standardization of cloud API’s, it would make the job of the integrator much easier because they could create a single API, write to it and have it work anywhere.

Is pricing a tough one to figure out too?

The cloud infrastructure market does not have a standard unit of computing power, which makes it tough to derive a proper pricing model. All cloud service providers differ in terms of their pricing. Each service provider assembles a compute package that's different enough from the others, thereby making it difficult to draw a comparison between the two.