All experts blogs

Why Artificial Intelligence is Not a Quantum Leap but Merely Makes Our Data-driven Work More Effective

Recording, interpreting and acting on data. Anything that artificial intelligence could one day do for us, we must first understand ourselves and be able to do on a small scale. We need cross-functional teams who are entirely data-driven in their work and who can rely on persistent and customer data that is as standardized as possible. We have a long way to go to achieve this.

Online Marketing is Becoming Increasingly Complex

Digital marketing teams operate across a wide range of disciplines. Owned media including websites, blogs and apps take center stage. Social, paid and earned media each play a role to varying degrees. Marketing channels continue to diverge and increase in complexity, not least because the limited attention available is subject to fierce competition and screens are shrinking in size. It is becoming more challenging for generalists to use a variety of marketing channels simultaneously and still have an impact. Outsourcing to agencies increases maturity in individual disciplines; however, since external experts tend to only interact with the client rather than forming a network, no real integration can take place.

The First Step in Data-driven Marketing

I often ask digital marketing teams whether their work in individual marketing channels is data-driven. They usually reply that it is. Everyone monitors and reports on a set of metrics. These include open rates, click-through rates, conversion rates and checkout rates, to name a few. I then frame the question a little differently: „If a metric changes, does this directly impact how you act? Does this raise additional questions that you then discuss with the people responsible for the channels and the business?“ Here, a „yes“ is usually followed by a number of concessions. „And if you have any questions about user interaction behavior, is the first thing you think of the target user journey from last year’s UX workshop?“ The nod makes it very clear that the team is still guided by a prospective and planning-oriented outlook.

For retrospective analysis, valid data is the more appropriate place to start. By using differentiated and segmented analyses of user behavior, facts and figures are collected that form the basis for hypothesizing and ongoing optimization. This is generally a sandbox for retargeting, conversion optimization and split tests. Targeted influence is exerted on a defined user segment with the aim of increasing a specific indicator. User segments group together users with similar prerequisites. For example: „currently browsing on a mobile device and/or has not made a purchase for 180 days“ or „has added products to the shopping cart but has not yet proceeded to checkout“ ...There are no limits to your creativity.

Why We Are Currently Focusing on Data Rather Than the Customer

Previous approaches have used cookie-based information. Although several visits can be aggregated into one journey, the data has not yet been linked over different platforms. This is why the marketing team often nurtures leads on one platform without taking into account how the same customers behave in the app or interact with the customer service team.

Even highly data-driven marketing teams often act in silos and have a fragmented view of the customer. Is this reasonable? It is, as the data-based optimization of the user journey is a cross-functional discipline that can even be put into practice on a small and imperfect scale. In fact, the reduced complexity leads to faster results and a steeper learning curve. In light of this, the organization should note that the proven uplift of a specific measure only represents a presumed improvement and familiarizes employees with the method.

It’s worth remembering that we will continue to misinterpret data and take wrong turns in the future – although hopefully on a higher level. As long as we continue to learn, we are on the right path. Only by learning can we acquire the basic understanding that will help us to develop the algorithms that will drive artificial intelligence in the future. Before we can train a machine to “behave intelligently”, we must have a firm grasp of the basics and understand how user flows work ourselves.

Andrea Malele derives the hypotheses for the optimized user experience from the data.
Andrea Malele derives the hypotheses for the optimized user experience from the data.
The fact that we do not have a complete overview of our customers’ data should not prevent us from acting today but should inspire the way we think about the future.

Why CRM is Not Suited to Being a Control Center

As described above, it is primarily a website or an app that generates user data. Some of this data is already stored in the organization’s Customer Relationship Management system as a prospect or customer. One example is the connection between the call center infrastructure and the CRM. Upon receiving a customer call, the agent sees the customer’s current subscriptions and the benefits the customer has already received in real time.

Based on contact history, lead nurturing can be planned and carried out in the form of marketing campaigns or sales offensives. For this purpose, sophisticated user segments are collated and passed on to the campaign tools. However, interactions in one channel will not influence how the campaign is playing out in another channel in real time. In the event that the conversion has already taken place via a newsletter, this will not directly affect the Facebook campaign. Although the CRM can obtain information from other systems or provide data to those systems, it is not fundamentally suited to real-time organization. This is because it only includes prospects and customers who have already progressed to the later stages of the conversion funnel.

The Unified Customer Perspective. A True Pot of Gold

As described above, every platform uses cookies to record user behavior over the course of a session. The profile remains anonymous until the user enters an email address into a web form or logs in. In this way, every platform maintains a dataset and adds to it over time, resulting in duplications and diverging customer profiles. Without a unified customer perspective, a personalized customer experience can never be truly achieved.

An excellent customer experience can only be provided by collating all available data on the customer over all systems and touchpoints – and doing this cross-platform, cross-device and cross-channel. This is what a „Customer Data Platform“ (CDP) does in its capacity as middleware.

How a „Customer Data Platform“ Closes the Gaps

Customer data is collected from different data sources and uploaded onto the Customer Data Platform or CDP via connectors. Here, the data engineer standardizes and transforms customer identities into a unified customer profile. The data is collated on the basis of a unique ID that is derived from an email address or a telephone number. The consistent database is the data scientist’s raw material. The data scientist explores, sorts, filters and searches data systematically for patterns and correlations. This results in intelligent segmentation. Segments are made available to all distribution channels via standard connectors in near time. The marketing teams are then able to tailor their campaigns more specifically to the needs of the individual target groups. Will this make everything quicker, simpler and more efficient? Far from it. The amount of effort required will increase with every additional segment but hopefully in lockstep with a rise in effectiveness.

The Martech Industry is Buzzing

The term „Customer Data Platform“ was coined by David Raab, founder and representative of the CDP Institute. Every six months, the Institute publishes a report that classifies providers. The most recent publication dated July 2019 lists and evaluates a total of 96 vendors, as opposed to 63 in the previous year. The Gartner Hype Cycle for Digital Marketing has also been confirming inflationary expectations in this area for three years now. The growth rates in the industry are enormous and the exploratory phase does not appear to be moving in the direction of consolidation any time soon.

Some major marketing cloud providers have now publicly announced their intention to integrate a CDP function into their suites in the near future. Wherever you look, buzzwords such as customer experience, 360° customer perspective and real-time marketing are coming into the spotlight. However, many suites are still not listed in the CDP Institute’s overview of providers.

A Customer Data Platform is a packaged software that creates a persistent, unified customer database that is accessible to other systems.

Why Denmark is Excelling at E-Government

Decades before CDP emerged on the marketing horizon, one country was already on its way there. Today, Denmark leads the world in the UN’s E-Government index. In 1968, a central database for the resident population of the country was launched. It then took a long time for the authorities to finally declare electronic invoicing obligatory for all suppliers in 2004, which naturally required a layer across government departments. Consequently, the digital channel was also introduced for private individuals and soon after, in 2007, the digital mailbox replaced the paper standard in G2B and G2C. This shows that standardized customer identification can trigger an enormous multiplier effect.

More On This Topic In Our Dossier On Artificial Intelligence

Efficient Image Management With AI

The formatting of images poses great challenges for content managers. AI can relieve this burden by automatically calculating the focal point. This ensures that the interesting image section always remains in view.