With more and more consumer data online, recommendation engines are becoming big business. Stephan Noller, CEO of nugg.ad and IAB Europe Chair of the Policy Committee, looks at how predictive technologies have altered the way we browse the web, and the implications for ecommerce and beyond.
Recommendation engines have drastically altered how people navigate the web, consume content and make decisions. Predictive algorithms, as they are otherwise known, are an incredibly useful tool but consumers are often unaware of how the technology functions and the added benefits it brings to their digital shopping experience. However, give them an example of the technology in action on platforms such as Amazon, a site which is well known for the usefulness of its suggested reads, and the advantages become clear.
Technological evolution has allowed recommendation engines to yield increasingly sophisticated results. Improvements in technology and programming means predictive algorithms can now make decisions at an astonishing 100,000 times per second during peak times. These decisions are responsible for determining who should be shown a particular advert or creative, as well as the timing and placement. For advertisers there’s no question that tools combining customer data with a predictive algorithm to make intelligent assumptions on purchasing behaviour, play a vital role in driving sales.
Such technology is an invaluable asset online, enabling advertisers to deliver highly targeted communications at consumers. Previously, shoppers relied on straightforward search to discover new information and it was much harder to discover new information, products or content.
In recent years, social media has enabled consumers to discover new content through sharing with friends. Shared content opens the doors to previously untracked corners of the internet, but the method for discovery remains limited. Friends usually have similar interests and consequently engage with similar content. However, predictive algorithms offer consumers an unlimited supply of new recommendations based on a breadth of interests. These algorithms process large amounts of data and consequently not all recommendations will be perfect. Until algorithms are programmed not for commercial objectives but human logic, and hence risk, recommendation engines will still be judged by their mistakes not their intuition. Nevertheless, the technology is continuously evolving and algorithm tools are learning with every click.
A major issue facing organisations today is the negativity surrounding recommendation engines, the key being that customers dislike thinking of themselves as predictable or too similar to someone else. Another major obstacle is the disconnect between the business agenda behind search engines and the customer’s expectations, alongside the trust issue. For algorithms to be viewed more positively, organisations need to be more transparent, by explaining the metrics behind the engine to their customers and providing the option to use the algorithm as the driver for search.
Moving forward, as the technology becomes more refined, algorithms may anticipate consumer behaviour. For example, an algorithm could predict whether someone might return a product purchased online through analysing past behaviour. As our lives become increasingly linked with the online world and more of our daily activities are organised via digital platforms, recommendation engines will play an increasingly useful role in our lives, helping us to choose food, books, music and where we shop. Advertisers that rely on influencing the consumers behaviour in the shop will find ways to make even that predictable and optimise their digital campaigns accordingly. Large brands will even allow consumers to decide how they’d prefer to receive recommendations. Organisations such as Amazon may adopt a scalable functionality covering serendipity (pot luck) to targeted recommendations based on consumer profiles. Innovative technology such as Apple’s Siri on the iPhone demonstrates the potential of the technology. Siri learns new things, analyses our speech, recognises tonality and generates an analysis of our favourite topics on huge server farms.
Rapid adoption of smartphone and tablet technology has significantly sped up the transition towards accurate and relevant online recommendations. Brands and advertisers can draw on the wealth of data gathered from mobile devices to inform and determine the kind of content that is transmitted to consumers. Location tracking introduces a particularly valuable new dimension to the algorithm equation. Organisations can now build an even more complex profile of consumers using mobile devices by pinpointing their location. This development makes the traditional recommendation engine determined by click-through’s seem antiquated in comparison.
Whilst data gathered through mobile devices and online is invaluable to the marketer, the Big Brother comparisons are inevitable. Therefore it’s crucial that advertisers implement a best practice policy around data privacy in order to drive consumer confidence. They must always protect consumer privacy if they want to continue collecting data that effectively fuels the algorithm. Companies should therefore inform customers of how their data is held and used in their privacy policy, make this information easily accessible as well as offer them possibilities to opt-out of any tracking, etc .
Anonymising customer data is one way companies can reassure consumers, as well as effectively use their data for targeting purposes so that the value-return for the consumer is obvious. This involves omitting personally identifiable data, such as name, address, email or IP address, with all predictions based on statistical models. By maintaining a clear privacy policy, companies will have fewer issues with opt-out and will build greater trust with customers.
Advertisers who fail to grasp the benefits of algorithms must do so now and unlock the potential for their brand, particularly as the technology is constantly evolving. At the same time, advertisers must be transparent about how they use customer data if they are to build trust and obtain access to the valuable data in the first
by Stephan Noller
CEO of nugg.ad and IAB Europe Chair of the Policy Committee
http://nugg.ad/en/
