In conversation with Saim Rolf Alkan


(Published in The Produktkulturmagazin issue 4 2016)

Online trade is on route to become a two-tier class society. Observers see a development in which Amazon, Otto and other big players dominate the markets with their capabilities and capital at hand. In conversation with Saim Rolf Alkan, Managing Director of ax semantics, a technology leader in automated copywriting, Temel Kahyaoglu, Chief Analyst at The Group of Analysts, discusses the new requirements for online trading and whether marketing automation holds the answer to future questions.

Mr. Alkan, what challenges do companies face in e-commerce?

I see a whole series of challenges. However, the process of concentration in the online trade overshadows everything. Major players dominate, and national as well as niche shops must change constantly in order to exist. Central challenges are process and cost optimisation, internationalisation and agility, which must be developed despite scarce resources. Automation will play a cardinal role. Above all, the willingness to look for potentials in the areas that were previously regarded as not automatable will create competitive advantages. Examples of this are marketing automation and automatic copywriting.

How have they changed in recent years?

If you allow, I would like to speak not of years but of months rather. The constant and accelerated change requires a higher agility and the willingness as well as the structures in the company to convert learnings much faster into successful actions. The online trade is currently divided into the hungry and the well-fed. An oversaturated condition that is unhealthy because it paralyses and blinds in the face of the development of the major player. By this, I mean not the willingness to invest but above all the lacking willingness to also evaluate and adjust those processes that function well supposedly. Therefore, much potential remains undetected. The reasons for this lay in corporate culture and leadership. Unfortunately, it is not self-evident to understand mistakes as a basis for improvement and development rather than to deduce responsibilities. This is an important task for managers and organisations.

How does content creation work at the push of a button? What are the benefits for a business?

With ax semantics, we provide a SAAS platform in which an online merchant’s copywriter automatically creates texts based on structured data. The tool for this is a text editor which is only slightly more complex in its functions than the creation of a serial letter. That is why just one day of training and one day of practice is required to be able to use this semantic software. It is ready to go after that. It is available in more than 15 languages at the moment and in a total of 42 languages on request. The SAAS platform has the advantage that all languages and other technical innovations, such as automatic data collection for product data, for example, are available at any one time on the customer site. This results in cost and, above all, speed advantages for the enterprises. Within a few seconds, texts are generated in several languages for many thousands of pages. However, I currently see the greatest advantage in Automated Conversion Optimisation. It surpasses all other advantages. Ax semantics measures the success of generated copy and then reproduces variants similar to the lucrative texts. This means that with every purchase in the online shop, insights from customer behaviour are inserted directly into the copy and the sales increase. Thus, online stores become learning and agile companies and insights are consciously used to grow.

What challenges do you see in the internationalisation of companies?

The subject of content is generally one of the biggest challenges in online trading. Especially in the context of an internationalisation strategy, it plays a role, if not the decisive one. In many countries, foreign shops without content in other languages are not listed at all. In Italy, a shop has to present Italian content in order to emerge in the most important price-cutting machines, for example. We are currently supporting a tire dealer with around 60,000 products abroad. The text machine that he configures creates product descriptions in German and two other languages - fully automatically for each stock and also for each new product – and optimises the conversion automatically. The cost of the text machine is about 2,000 euros per month. For this amount, neither a single individual text nor a translation could be financed. The constant self-optimisation of the copy is not taken into account in this business case but would further improve the ROI.

What are examples of companies or groups that have successfully implemented internationalisation? What have they done right in your eyes?

Through the international lens, I see Amazon as an outstanding example in e-commerce. Unlike Zappos, Amazon has never been content with the American domestic market. Globalisation, the creation of new business segments and the constant desire for improvement have made this major player the benchmark for online trading.

How do you see Germany in comparison?

In Germany, I see no wide online retailer with a comparable broad strategy. Certainly, Zalando is one of the most successful examples in Germany, and there are similar strategies for internationalisation as well, but overall there is still a lot of catching up to be done in German e-commerce. The EU offers an internal market comparable to the U.S., but it needs to be served in several languages. The desire for automation is still very small in marketing departments and the principle of “fail to learn” is often incompatible with corporate cultures.

How can text automation be utilised in globalisation?

In concrete terms: it needs well-structured data from PIM or WaWi, or from CRM or Analytics. All of this together is controlled by text generation, taking into account localisation, such as cultural or legal aspects and the tonality determined by marketing. This means country-specific content and, if CRM data is integrated, even individualised content in a language coloration corresponding to the brand. Through the automatic generation, the grammar is applied in the same quality during the creation of the copy, thus creating a consistent quality with regard to spelling, completeness and tonality superior to the human author. To understand this quality jump, it is important to understand copy generation. The text machine does not translate but recreates every text according to linguistic and semantic rules. This results in error-free international content and is a giant step towards globalisation for online retailers.

What role will artificial intelligence play in the future of e-commerce?

AI, Machine Learning, Big Data - all these describe the future of e-commerce when it comes to the individual shopping experience. In the stationary trade, successful sellers classify clients, are empathic and carry out their requirements. This can also happen in online trading. Not with the methods of man but with the help of AI, and we thus increase the positive shopping experience of each consumer. We are already discussing such initial approaches with clients. For example, with questions, we create customer-specific demand profiles and subsequently formulate product descriptions of suitable merchandise for the individual who then feels understood and sees this understanding reflected in product and copy. Different methods, but the results are comparable to the sales consultant in the stationary trade.

Contextual commerce is seen as one of the most important trends in the coming year. What do you think?

Absolutely. The future lies outside the internet. Voice control, dialogue and the IoT are the new contact points. Based on my previous order behaviour and household inventory, text machines will soon be producing spoken or visualised content in order to make buying recommendations. In addition, the evaluation of data from social media will help to present new products in the smallest target groups. What sounds like a distant future today, is to begin tomorrow.

Loved this article? Put it in your cart!


ax semantics is a software that interprets data and thus generates texts in up to 15 languages. The configuration language ATML3 allows users to train and optimise the artificial intelligence themselves.

ax semantics

Picture credit ©Philipp Renger


Leave a comment

Please note, comments must be approved before they are published