Best Practice in Marketing – Management Focus Group „Big Data and the Digital Revolution – Integrating Clicks & Bricks“

7 strategic insights for the successful implementation of Big Data in your company – based on insights of the Management Focus Group „Big Data and the Digital Revolution – Integrating Clicks & Bricks“ (September 29th/30th, 2015 in Berlin)

Insight 1: Disruptive innovations cannot be made with the help of Big Data

As Prof. Dr. Marcus Schögel explained, disruptive innovations often completely displace existing technologies, products or services. A disruptive innovation usually creates a fundamentally new market and a new value network. Moreover, disruptive innovations stand in contrast to sustaining innovations which can take the form of discontinuous or continuous innovations.

The Innovator's Dilemma (Christensen 1997)
The Innovator’s Dilemma (Christensen 1997)

An important difference between sustaining and disruptive innovations is that sustaining innovations concern the technology while disruptive innovations cause fundamental market changes. The challenge with disruptive innovations is that they are usually not profitable in the beginning. Furthermore, when investing in disruptive technologies, there might not be enough money left to make investments into sustaining innovations.

However, when striving for disruptive innovations, you shouldn’t rely on Big Data analyses. Such analyses are only useful when it comes to support the development of sustaining innovations because they rely on historical data and cannot really foresee the development of unexpected products or services. Go beyond the hype and use Big Data usefully when it comes to product development.

Insight 2: Rethink your advertising strategy

As mobile search and shopping happens all day, it is crucial that data gained from mobile searches are being interpreted properly. Always keep in mind that search engine optimization is important when it comes to internet advertising, as almost nobody looks at the Google search results that do not appear on the front side. So if you like to hide a corpse, place it on page 2 of the Google search results.

At Ricardo, one of the leading Swiss internet auction houses, they follow a particular approach in online advertising. They test the conversion rate of different subjects in online campaigns and adapt it accordingly. However, by following this revolutionary approach, Ricardo produces a lot of campaign material that is not being used in the end. That is one of the reasons why they decided to develop the concepts of the campaigns in-house instead of outsourcing them.

Furthermore, Ricardo successfully implemented an email-newsletter that is completely personalized (targeting the segment of one). They programmed an algorithm that uses Big Data to calculate the perceived similarity of products. In the email-newsletter, they refer back to auction opportunities the customer might have been interested in earlier on. With this personalized newsletter, Ricardo succeeds in reactivating customers by calling their attention to potentially interesting auctions and can lift the rate of how many newsletters are being opened by the receivers. This finally has a substantial influence on the company’s performance.

Insight 3: Big Data has to be transformed into Smart Data

Big Data analyses are a demanding task. It is definitely not easy to get valuable insights from Big Data. In order to carry out good Big Data analyses, you need to build up a team that consists of external and internal employees. However, it is crucial that you establish a certain in-house-competence in the field of Big Data analysis so that you don’t exclusively have to rely on external consultants.

Furthermore, it is recommended to find a “sponsor” within the company to get the necessary support for your Big Data projects. Having said that, it is also necessary to create check lists on which you define your goals of research. However, always keep in mind data privacy regulations as well as your digital media budget. Moreover, a feasibility study may help you to check whether the pursued insights can be implemented at all.

GfK (2015)
GfK (2015)

Besides that, it is vitally important to strive for inter-divisional cooperation (e.g. with the departments IT, CRM, procurement, agencies, vendors etc.). However, this might be a very challenging task as in many companies the departments think and act in isolated silos.

What is more, it is essential to take enough time to present the results of the Big Data analysis to the top management in order to get their support. Finally, when it comes to implementation, look for quick wins to keep the motivation of your employees high and to show top management that your strategy works.

To sum up, one can say that Big Data is an important issue regarding the generation of insights for business. However, there is a take-away that you should keep in mind when dealing with Big Data: You shouldn’t necessarily rely on Big Data only. If you do so, then it is crucial to understand the consumer context and to combine consumer data with “reference” data in order to gain better insights.

Insight 4: Most decisions can be automated

Automated decisions make business sense because they are faster, cheaper and better than human ones. For example, the processing capacity of a computer doubles every 18 months while the price of it stays roughly the same. Automated decisions play an important role, especially in industries with a low average profit margin such as retail, for example. When decisions are automated, companies can avoid that managers take wrong decisions because they trap into a decision bias, for example in sales planning, logistics or order quantities. At the same time, managers have more time for taking strategic decisions that cannot be automated. However, computers are better at solving complex optimizing problems. It is therefore important for companies to understand that humans and computers are good at different activities. The best approach is to combine the strengths of both of them.

However, the handling of data is not as easy as it seems. Data not only need to be collected but also to be stored and analyzed. This is the prerequisite for getting reliable predictions about the future and for making decisions out of the collected data. In the end, this must be the goal of collecting data at all, namely to take decisions based on (Big) Data.

“Maslow Pyramid” of Data Processing (Blue Yonder 2015)

99.9% of all business decisions can be automated, Lars Trieloff from Blue Yonder believes. However, there remain some decisions that cannot be outsourced to computers. For example, computers cannot help with the decision whether or not to enter a new geographical market. For taking this decision, human intuition and experience is needed. However, computers are very useful when it comes to making decisions in the daily operations of a retailer.

But why is it furthermore useful to delegate decisions to computers? In 90% of the cases where it would be necessary, no decision is being made at all. Still, this doesn’t mean that you don’t have to decide anything because making no decision is also a decision. It simply means that you decide to do nothing. However, although sometimes doing nothing might be a good strategy, it usually only works when your company disposes of a large return on sales.

Furthermore, many decisions are made by relying on business rules. The problem here is that such business rules can be contradictory, incomplete and complex. However, business rules simplify your daily work in a tricky way. By applying them, you don’t have to ask yourself anymore if what you do is really relevant. This can lead to decisions that are far from ideal.

In the end, it is not a question whether to rely on decisions made by computers or humans but to combine both of them in a smart way. That is also due to the fact that computers often struggle with tasks that are easy to conduct for humans. For example, computers are good at recognizing human faces but they struggle when it comes to identify abstract objects.

Another reason why it often makes sense to outsource decisions to computers is the fact that own bad decisions are usually attributed to the context, while mistakes of someone else are being attributed to the characteristics of this person. This misperception is called fundamental attribution error.

Insight 5: Targeting your customers purposefully is essential

Making successful business is often about finding the most valuable customers and targeting them purposefully. However, not every potential customer likes it to be targeted. According to the uplift modeling, the so-called “sleeping dogs” stop buying when you target them. However, there is one customer group that companies are interested in targeting. These are the “persuadables” which only buy if they get targeted. There is also another customer group, called the neutral customers. They either buy in any case (“sure things”) or they don’t buy anyway (“lost causes”) irrespective of whether or not they are being targeted.

Uplift Model (Blue Yonder 2015)
Uplift Model (Blue Yonder 2015)

One way to further optimize returns is by reducing risk and cost as this increases revenue and profit. That is why it is essential to collect data to be able to make some trend estimations. Only then it is possible to classify your customers and to separate profitable from non-profitable ones. Like this, you are able to anticipate certain events and to take decisions purposefully.

Insight 6: Offline and online channels are no contradiction, they rather complement each other

Today, for many companies it has become difficult to differentiate themselves on a product level. This is especially true for the automobile industry where most markets are saturated. That is one of the reasons why Audi tries to differentiate itself from competitors by an innovative show room concept that has been established in London, Beijing and Berlin: the Audi City. With this show room concept, Audi tries to strengthen its premium brand positioning. The revolutionary store concept has been developed in consideration of the four megatrends urbanization, individualization, digitalization and the ever growing product portfolio.

Audi City Berlin (2015)
Audi City Berlin (2015)

With the Audi City, the car maker ensures the Audi brand presence in the city center with a representative flagship store. However, not the exhibited cars are in the focus in the Audi City. Audi rather pursues a completely new level of customer approach and interaction by integrating concepts from the digital world. On the ground floor, the customer can get advice from Audi guides. Those guides help the potential customer to find its way in the shop and to configure the dream car at one of the multimedia touchscreens located in the shop. The customer has furthermore the possibility to print out the individually configured car documentation and to take it home. Those customers who wish a more personalized sales advice are being escorted to the first floor where an Audi Expert supports the buying process further on.

Insight 7: Don’t cherish the ‘turkey cock illusion’

Big Data analyses are useful when you want to analyze things that happened in the past. However, they are not a good approach when making predictions for the future. The following example of Prof. Dr. Gerd Gigerenzer illustrates this in an impressive way.

Many businesses are dependent on the dollar-euro exchange rate. Therefore, banks predict the exchange rates for the following year. However, even if these predictions are made by professional analysts, they are usually wrong. This is due to the fact that the experts assume that next year’s trend will be the same as in the previous year. If you analyze the predictions you find that in most cases, the forecasts of the experts follow the trend of the previous year.

At this point you may ask yourself why such predictions are being made at all if they are wrong anyway. The reason is that managers of big corporations want to rely on these predictions so that they have someone to blame when they take wrong decisions based on the misestimated exchange rate. To simplify, you can say that only due to the fear of managers to take personal responsibility such worthless analyses are being conducted at all.

This example is also known under the notion “turkey cock illusion” which describes the case of a turkey cock that is being fed by a man. First, the turkey cock is afraid of the man but as soon as he realizes that the man doesn’t hurt him, he starts to trust him. The man continues feeding the turkey cock which becomes more trusting every day. However, when the day of Thanksgiving approaches, the man doesn’t feed the turkey cock anymore but kills him. This happened due to the fact that the turkey cock expected what would happen in the future by analyzing what has happened in the past.