Skip to main content

Digital marketing’s data dependency

(Image credit: Image source: Shutterstock/Maksim Kabakou)

Digital marketing has always been dependent on data, and as competition heats up in 2020, marketers will seek to differentiate themselves through the data and audience insights they have to hand.

According to Altimer, over half of marketers agree that using more data to create customised material to aid the customer journey is a top priority in the next twelve months. Additionally, new data sets will continue to act as an informer to overarching marketing strategy, as well as customer experiences.

Advancements in artificial intelligence (AI) and machine learning (ML) will facilitate a resurgent interest in data on customer trends and habits. The processes of collating this data will become increasingly automated, saving valuable time and money that can be spent on value-added activity.

This will enable marketers to reach customers at the right touchpoints at the right time via the right channels. As a result, marketers will also gain an understanding of what purchases their audiences are considering through implicit clues, and use that information to influence ongoing planning thereafter.

Exploring different channels

The best digital marketers distinguish themselves by making products that appeal to individual customers. Epsilon found that customers are 80 per cent more likely to purchase a product or service from a brand that provides a personalised experience. Critical to this is targeting the right customer via the right channel.

Previously customers have been segmented to allow for the tailoring of impactful marketing strategies. One-to-one marketing takes this a step further, applying the same concept to the individual, and AI turns this idea into a scalable reality. Processes which are mundane when conducted manually, such as examining and collating data into a profile, become automated. AI has already been implemented in various channels. Google Ads offers automated bidding to produce an optimum strategy to maximise impressions, clicks or other metrics.

These applications are currently channel specific; 2020 will see the emergence of more omnichannel platforms. Tracking and tailoring activity across a range of platforms would allow for a more comprehensive and personalised experience for each customer. This enables delivery of the most relevant content through a customer’s preferred channel. The results are undeniable: 88 per cent of marketers reported seeing measurable improvements after implementing a personalisation strategy.

Conversational analysis

Candid conversation is used in sociological studies to understand and predict human behaviour; the same approach has now been applied to data science. Speech has long been an established indicator of behaviour, but the digital age has seen a relocation of our conversations to social media. Now conversational analysis uses machine learning tools to scan millions of statements and discussions across social media and recommend certain products and services automatically.

Social listening tools aren’t new, but 2020 will see them become more sophisticated. For the first time conversational analysis will enable data scientists to glean implicit as well as explicit sentiment. For instance, an enthusiastic Facebook update along the lines of “it’s so exciting to finally be able to decorate our home!”, would previously have flagged up as an opportunity for an individual to be targeted with furniture and paint adverts. The emerging updates of conversational analysis tools will infer from the excited tone that this individual is a first-time buyer and will definitely want to take care of their new place. Products like building and contents insurance will then be pushed to fulfil this subliminal brief.

These tools will be invaluable to marketers and advertisers, enabling them able to graduate beyond broad speculation of what people may want to buy, to understand how each active person feels about certain products. If used wisely, we will be able to anticipate the purchase patterns of individuals.

Predictive marketing

Predictive marketing has emerged as the new paradigm in data science that marketers are looking to take advantage of. Conceptually, it attempts to extract insights from large datasets in order to predict future outcomes, with greater efficiency and higher levels of automation. Although still in its nascent stages, predictive models hold a great deal of potential to streamline marketing processes. The technique uses historical data and reinforced learning to determine the probability that a marketing strategy will succeed. 

Reinforcement learning means the agent can develop and improve marketing decisions in a continuous informative process. Depending on the outcome, each action in the environment is either associated with a reward or a punishment and the agent learns what actions are beneficial to repeat in a particular scenario. The design of a reinforcement model can evolve around any KPI as a metric, allowing the marketer to tailor the automated strategy to the client’s needs.

Deep learning techniques will increase in significance in 2020. By automating marketing processes copious amounts of time will be saved. Adjusting the parameters of models to suit a specific KPI is an extremely innovative and handy feature and allows for traceable results.

Margins of competition between agencies are become slimmer. In 2020, they will seek to differentiate themselves regarding different KPIs. Innovations in AI and ML will become increasingly popular to gain as much insight into individual customers as possible. As such, the ability of data and metrics to create a competitive edge should not be ignored.

Andreas Pourous, CEO & Cofounder, Greenlight Digital