Marketing data analysis with IBM-Watson fails2019-09-04T11:00:21+01:00

Marketing automation and artificial intelligence with IBM Watson not possible

Questions and answers about IBM Watson in the area of marketing automation

The customer order was actually clearly defined: dynamic assignment of potential customer profiles to target groups in real time. Watson should independently weight the relevant parameters for the target group allocation. In addition, Watson should also find the right parameters for customer-to-target group allocation. Another requirement was to integrate the target groups into the customer’s Dynamic Pricing System.

To cut a long story short: We finally implemented the solution with Keras, the successor to TensorFlow from the Google Brain project. IBM-Watson could not cover the requirements. This was a reason for us to discuss some general questions about IBM-Watson within HighPots.
There are more and more doubts about Watson’s effectiveness. Project insiders from various companies are reporting on high costs with little benefits. There are also reservations regarding the security of data – keyword Cloud. Has the IBM-Watson project failed with regard to the cooperation between the marketing departments of the Fortune 2000 companies even before it had started properly?


Watson is a high-performance application that helps companies do their day-to-day business. Watson is significantly faster than the software solutions that most companies currently use. In this respect, Watson increases efficiencies. So far, everything seems to be in order. According to the project managers who drove the migration of Watson in the companies, there are, however, very unsatisfactory points.

point 1:

As technology currently stands, IBM Watson is not a market superior AI, although AI capabilities would need to be in place to justify its high price; the “IQ” of intelligence is disproportionate to the costs Watson entails – so companies fail to achieve their ROI in the necessary time frames.

point 2:

Watson’s setup, configuration and customizing effort is enormous (and not very intuitive). This affects many areas within the Watson application. For example, data protection departments that want to pseudonymize and anonymize data according to the new EU Data Protection Directive as well as scientists who want to simulate and extrapolate complex marketing target area analyses. Uncertainties therefore exist not only among gigantic corporations, but also among traditional, established medium-sized companies.

point 3:

The statement “First IBM makes a mainframe leasing contract with us, which costs millions per year in the usual manner, in order to next take millions from us for a solution that hardly runs on the mainframe anymore” is also frequently heard from decision makers.

Conclusion from question 1) – Doubts about Watson’s performance in marketing automation

But it is still too early to declare the cooperation between companies and IBM a failure. Nevertheless, IBM’s enterprise customers will examine IBM’s solutions and prices more closely in the future. Confidence in the IBM brand has declined across industries. As a result, IBM’s brand has become less competitive and will have to again compete in price wars with suppliers less well known in the industry. These are great opportunities for smaller companies, some of which have excellent solutions with a very good price/performance ratio, and also for other data-driven companies in Silicon Valley. Amazon, for example, is currently also investing in industries that were previously reserved for IBM – the current situation between IBM and its customers will certainly not harm Amazon.
Employees at IBM themselves are currently talking on the quiet about the ” too progressive” internal marketing Watson would have advertised too early as AI.

Question 2) IBM’s aspirations with Watson

IBM is hoping for good business in cooperation with Watson, especially in terms with top companies. When it comes to cooperation with insurance companies and banks, for example, IBM can look back on a long history; many insurers have been loyal users of IBM mainframe computers for decades. But where do you think the biggest difficulties currently lie with Watson?

Response Business Model IBM Watson:

Most insurance companies that use mainframes have long-term leasing and maintenance contracts with IBM. Communication in this area will therefore not change.
IBM’s business model was, among other things, to bring companies into the IBM cloud. Watson, for example, was the sales argument for switching to the cloud. The companies that are gradually migrating their applications to the IBM cloud therefore need less computing power year after year and as a consequence will no longer need high-performance mainframes at a certain point. The reduction in the mainframe business should therefore be set against the additional revenue in the cloud business – that was probably the way how IBM calculated its costs. But if the cloud business doesn’t start because Watson doesn’t serve as a reliable argument, that calculation won’t work. Blockchain could also develop into a serious threat for many of IBM’s business areas. After all, what would happen if insurance companies no longer needed a data center, and certainly no mainframe at all, due to Blockchain (e.g. in the areas of smart contracting)?

Question 3) How important was marketing for Watson?

Answer IBM Watson Marketing:

According to many IBM salespeople, Watson’s sales were supported by ingenious marketing. These salespeople sold Watson in good faith in their own marketing messages. But they were surprised by the negative news themselves and ended up dissatisfied with the marketing and overall positioning of the product. According to statements by IBM sales staff, however, internal measures are already underway to ” rearrange ” IBM sales. IBM must now invest millions in marketing and sales in order to make up for the damage; the negative Google search engine results and the criticism in social networks must be pushed to the back ranks by positive marketing messages. The increased demands on marketing are therefore very high as well as extremely vital.

Question 4) the greatest potentials of IBM Watson

Where do you see the greatest potential of IBM-Watson for companies? Is it too early to turn away from IBM-Watson as an IBM corporate customer?

Answer Potential IBM-Watson

IBM Watson corporate customers should have rethought their own loyalty to IBM long ago, not just since Watson. Watson is just a good example that loyalty is not appropriate in business. Many IBM competitors have been offering comparable solutions at lower prices for many years in numerous business areas. In the age of web-based, location-independent applications and services, IBM’s sales argument is also losing importance with regard to the high level of integration and interlocking of IBM products. Most of the companies we work with at HighPots use between 10% and 15% of the IBM product integration possibilities. Most of the product integrations that IBM’s corporate customers have lived with could easily have been developed by themselves without the purchase of expensive interface modules. In many cases, deep, rigid service interlocks actually are disadvantageous.