May 15, 2014
As Yahoo CEO Marissa Mayer attempts to return the company to its former market position, the story of how her company missed the opportunity to buy Google for $5 billion in 2002 must be on her mind (of course, Ms. Mayer was at Google at the time).
This post is not about missed opportunity, which happens quite frequently in technology. It is about why proprietary technology matters in mobile marketing.
Source of issues
Though companies can succeed in advertising technology because of scale, first-mover advantage or even having the best sales organization, only proprietary technology can provide a sustainable competitive advantage.
In ad tech, no company stands out more for its proprietary technology than Google.
From the development of the PageRank Search technology to technologies that Google acquired such as the Android operating system and DoubleClick, Google has placed the continued development of proprietary technology at the core of its roadmap for the company’s portfolio of advertising solutions. That is why Google products keep getting better and more sophisticated.
Unfortunately, too many of the companies in mobile ad technology rely on outsourced or open source technology. And even many of the companies which originally had proprietary technology have not continued developing it to stay ahead of advancements in the industry.
It is no wonder that mobile ad rates remain so low despite the strong growth in application and mobile Web usage.
Set on data
For example, consider mobile ad targeting technology.
Since the late 1990s, the vision for mobile marketing has been based on offering the right offer to the right user at the right time and right location – offering coffee lovers a coupon to a nearby Starbucks location during slower afternoon or evening hours.
Unfortunately, relevancy is still lacking in most of the mobile ads that I receive, despite the wealth of data available about my mobile habits.Why? Because too many mobile campaigns rely on simplistic targeting such as age group, phone type and general location. And then users are targeted according to these broad segments.
Mobile companies using advanced technology have the ability to make laser-like images of their data according to hundreds of data points and can then target ads based on the most relevant user response patterns, not segments, to increase conversions, while reducing unwanted clicks.
Success in ad targeting comes down to uncovering behavioral patterns which result in conversions.
Therefore, mobile ad technology companies need to be analyzing all of the available behavioral data points – and all mobile networks have access to trillions of data points – using statistical tools such as support vector machines (supervised/semi-supervised learning model algorithms) that analyze data and recognize patterns which are then used for regression analysis and classification.
Some patterns do not require advanced algorithms to uncover. We all understand that Coca-Cola will sell a lot of soda on a hot summer day.
But for a great portion of human behavior, the buying patterns are not so clear. The data is what we call heteroscedastic – data with differing variance which is frequently accompanied by a lot of volatility and fluctuation.
Therefore, to effectively analyze behavioristic data sets, marketers employ multilevel logistic regression to polytomous data, or data divided into secondary parts or branches.
THE AD TECHNOLOGY industry is suffering from low mobile marketing margins and revenue, and the solution will come from better technology.
Though this might be a mobile industry problem today, with the migration to mobile-first across devices and platforms, it will be a problem for the entire advertising industry very soon.
We need to place a greater mobile industry focus on creating better and more efficient mobile advertising technology to provide more value to mobile marketers, agencies, publishers and users to improve the economic model for the industry.