The Rise of Digital DNA

Digital DNA, or the idea that online activity leaves “footprints” containing information behind, is a hot topic for debate at the moment. Many experts agree that digital DNA can be used to predict behaviour and preferences, making it an ideal tool for marketing strategy. Others are cognisant of the dangers that digital DNA can pose, with the Cambridge Analytica scandal, where data was used unfairly and without peoples’ knowledge or consent, providing the perfect example.

 

This article will focus on the rise of digital DNA and consider how it can be used to help businesses achieve their marketing goals.  

 

 

 

A criminal at a crime scene might leave some of their DNA behind them, for example skin cells, fingerprints or footprints. The authorities can then gather this material and use it to trace the identity of the criminal. Digital DNA draws comparisons to this process, because when a user interacts with websites, videos, ads and other material on the internet, analysts can use the data that is left behind to identify what kind of person that user is, what their interests and preferences are and use this to predict how that individual is likely to behave in the future. Many retailers are now preserving information about how users interact with the internet and social media to do just that – predict how they are likely to behave or be influenced in the future. This can help to sell products and services. It can also help retailers to tailor marketing strategy to target people who are more likely to be interested in the products and services they wish to sell.

 

A good illustration of this is the rebranding of Toys R Us. Toys R Us was a childrens’ toy manufacturer, which was declared bankrupt in 2018. It rebranded as Tru Kids in 2019. As part of the rebrand, much of the Toys R Us infrastructure was preserved, for example some shops and some relationships with other brands and partners. Also preserved were large and complex datasets pertaining to their old customers, and the interaction of their old customers with the old Toys R Us brand. This data formed a central part of the new Tru Kids marketing strategy, because it was capable of being analysed and applied within a newly formulated marketing strategy for the new brand. As such, it might be said that the old brand’s “digital DNA” was used to assist the creation of the new Tru Kids brand.

 

Another good illustration is the Cambridge Analytica controversy. Cambridge Analytica was a data mining and data analysis firm, in operation between 2013-2018. It worked on the Leave.EU campaign, as well as on Donald Trump’s presidential campaign. In 2018, news broke that it had used personal information unethically, and additionally had obtained personal information without appropriate permission or even the prior knowledge of the user. In late 2018, the Information Commissioner was granted a warrant by a UK court to search the Cambridge Analytica server so that a full investigation could be carried out into the scope of their data handling. In 2018 the New York Times and the Observer reported that Cambridge Analytica had obtained and used information from Facebook that had been gathered for an academic project. This information related to the digital activity and preferences of up to 87 million Facebook users. Some have suggested that this data was then used to attempt to influence elections, for example through the targeting of certain “fake news” media campaign messages. The idea was that different people react differently to different types of “fake news”, or material, and the information that was obtained unethically was capable of being used to identify segments or groups of people, who were more likely to be receptive to certain messages. This concept of people reacting differently to different types of information was pioneered by Michal Kosinski in 2008 when he developed a basic profiling system, based on data including how people reacted to Facebook posts. Using this data Kosinski was able to accurately predict people’s behaviour. The profiling system was able to make more accurate predictions, compared to predictions made by friends and family of the user. Some suggest that subtly influencing registered voters, without their prior knowledge, in the way Cambridge Analytica was alleged to have done was unethical.

 

What is important to note is that data itself is not always useful. It must be interpreted or “cleaned” before it can be of value. This is why experts commenting on digital DNA have made the important distinction between technology use and technology alignment. Not all IT solutions are suitable for a given business, so successful digital DNA will involve an initial identification of what technology will deliver the best value and overall impact to the retailer proposing to use it. This approach of technology alignment is beneficial because technology is an expensive investment for any business. Furthermore, it can attract high costs, for example the costs surrounding security, data processing training and data protection. This makes a considered data strategy, which assesses which solutions are most beneficial, all the more important, because implementing a digital DNA strategy will often involve a digital “transformation” whereby appropriate questions are asked so that data capture of relevant data can begin. This enables a business to address specific nuances of their business and their operations with a personalised data driven strategy.

 

 

 

Digital DNA can be used to make processes including customer service, order fulfilment and supply chain processes much more efficient. The idea is simply an application of the ideas that underpinned the Cambridge Analytica controversy and indeed the Kosinski research, with the exception that the data used is obtained ethically.

 

Digital DNA can be used to identify business problems to begin with. Historical data about customer complaints, or negative reviews for example can be used to identify problem areas for a retailer. Once a pattern has been identified, human executives can then look at the problems and work to find a solution. For the digital DNA model to be effective, the correct actions need to be taken, based on the analysis of the data. A good example of how data was used to identify problem areas in a business setting is where digital DNA was applied to a water processing facility. Water processing is plagued by problems relating to floods, leaks, high and low water pressure, extreme weather conditions and interruptions to the flow of water. Data analytics was applied to the problem, and the water processing plant was able to deal with problems caused by heavy rainfall, before the problems caused issues like leaks and overflow by diverting water flow when this was necessitated by the weather conditions. This led to fewer leaks and overflows and less leaking and flooding. A combination of remote sensors and analysis of water pressure and flow enabled the plant to identify leaks and faulty pipes more quickly, ensuring that problems were fixed much more quickly, and in many cases before they caused major problems. The plant also connected their pumps to the internet and this enabled the water flow to be controlled remotely and much more effectively. These so-called smart pumps used machine learning to ensure the water was processed more efficiently.

 

Machine learning is defined as any system that analyses its own inputs, so that its outputs can be systematically improved. Historically, computers were programmed to perform certain tasks. However, with machine learning the machine is continually tweaking its own operation, based on analysis of data. Smart machines identify patterns in the data available, and the data is used to make decisions about how to improve efficiency or overall functioning. Using machine learning in the water processing plant helped the plant to understand where their historical problems were occurring and why. Information on past water leaks, overflows or suboptimal water pressure was used to divert water flow, thereby preventing past problems from being repeated.  

 

There are hundreds of applications of machine learning in business. For example, Walmart has applied machine learning and AI to integrate their sales data, with their supply chain management strategy. Data in real time is gathered at the point of sale and analysed. This data is then fed into their supply chain management strategy. This enables the retailer to predict what items are selling better than others, and which items are in highest demand. Then, the data can be used to ensure that stock is replenished efficiently, reducing instances of stock being designated “out of stock”.

 

This ability to prevent being “out of stock” makes a massive difference to overall profitability. An example is REI, a larger retailer in the outdoor clothing sector. Prior to their introduction of an order management optimisation tool REI reported that in the region of 800,000 sales were lost due to “out of stock” status. The stock optimisation tool used machine learning and AI to predict how many orders required fulfilment at given times and the store moved some of its supply chain to physical stores, so that some orders could be fulfilled on the same day. This strategy virtually eliminated REI’s problem with order fulfilment. A significant boost in profits was also reported as a direct result of the new order fulfilment strategy.

 

Out of stock status doesn’t just affect retailer profitability. There is compelling evidence that an out of stock status can also have a significant impact on brand recognition, with brands frequently reporting out of stock status seen as less reliable and less convenient. Research commissioned in 2018 by Which? aimed to identify the best and worst online retailers for 2018, and their survey featured thousands of retailers and more than 10,000 customers. Which? took account of a number of factors in its assessment of the best and worst online retailers, and customer service and convenience featured highly in the criteria used to decide which brands were performing better. Homebase was selected as the worst online retailer for 2018, and one of the central factors in this assessment was continuous out of stock notifications for items that were displayed on the website as “for sale”, only for closer examination to reveal that the item was actually listed as out of stock. Other concerns expressed by customers who rated Homebase in this survey included poor value for money and poor quality of items sold.

 

The Which? survey specifically addressed how customers interacted with a retailers’ website and from this research factors like product range and how user friendly a website was including how fast it was were taken account of in the quest to identify the best and the worst retailers of 2018. Similar research by Which? in 2019, revealed Ryanair as the worst retailer, with online bank First Direct getting the coveted first place. The bottom five retailers were Virgin Media, BT, Scottish Power and TalkTalk, whereas the top five positions were taken by First Direct, Waitrose, M & S, Lakeland and Waterstone’s. As in 2018, customer service was central to the selection of best and worst retailers. Ryanair has struggled in recent years with continuous reports in the press regarding poor customer service, delays and expensive prices for additional items like extra space and additional baggage. The 2019 survey asked over 4000 customers about their experiences and the survey specifically addressed how a brand made the customer feel, how well complaints were handled and how knowledgeable staff were. In the case of Ryanair, more than half of customers surveyed on how well Ryanair handled complaints placed their experience in the lowest possible category. Levels of trustworthiness also featured in the selection of the best and worst retailers, with Ryanair receiving heavy criticism for extra charges that were perceived as “hidden”, or “sneaky”. One respondent commented “Ryanair seems to make things deliberately difficult for customers in order to make more money out of its customers”.

 

The good news for retailers is that AI can be adapted to address every single one of these problem areas, including how well customers can interact with a website and how user-friendly it is. The correct data collection and data analysis strategy can quickly and conveniently enable retailers to see where they are performing poorly and enact a plan to remedy this. Chatbots and voice interactive technology allows for more flexible interaction between user and the website, and these also enhance the customer experience of using a website.

 

Digital DNA also has applications within targeted marketing campaigns. A good example is Target and its pregnancy discounts, which involved Target using point of sale data to identify customers who were likely to be purchasing pregnancy related products in the next few weeks. Targeted marketing campaigns followed, and these were extremely successful, perhaps, owing to how uniquely and precisely targeted they were.

 

Further, digital DNA can be coordinated with promotions to drive more sales and in particular to drive sales and footfall at times when a retailer is not busy. Digital DNA can be analysed to identify which cleavages and social segments are or aren’t engaging with the brand, and strategies can be devised and put in place to address any problems.

 

Digital DNA can be used to identify what types of sales are being made and what channels are being used the most for delivery. A retailer may find that this analysis allows them to work hard to capture more sales within a particular sales category, like digital sales or sales from brick and mortar stores. Equally, when a retailer has a good understanding of what types of customers they are attracting and what preferences and behaviour they exhibit, this information can be invaluable to the retailer’s marketing strategy. A good example is Starbucks who with digital DNA were able to identify a phenomenon dubbed “the Starbucks golden goose”. Starbucks analysis of its historical sales data enabled it to see that one flavour – pumpkin spice was selling particularly well. Using this information the brand made strategic decisions to expand its product range surrounding that popular flavour. This enabled it to benefit from unprecedented growth and profits.

 

 

 

Digital DNA is being used more widely because there is compelling evidence that it works well to help retailers and businesses identify ways in which to improve their operations within a plethora of areas where waste and inefficiency can thwart business profitability including supply chains, promotions, website functionality, order fulfilment and stock management.

 

As we have seen, a good digital DNA strategy focuses on the individual business and the individual problems that are faced. A business may wish to improve its overall sales performance through implementation of a digital DNA strategy, whereas some strategies many be more reactive and focused on identifying and defining particular problems so they can be fully understood and addressed. Many businesses, especially smaller businesses use “guesswork” to figure out what is wrong when sales aren’t as expected or when a strategy doesn’t work out as planned, but this invites a significant margin for error, even for seasoned professionals with years of experience. One of the principal applications of digital DNA is that it can reduce this scope for error and force a business or organisation to focus singularly on what is important and what needs to be worked on, or improved.