68% of retailers found ratings and reviews to be a most effective tool for increasing sales. Forrester

62% of all U.S. consumers read reviews before buying. lightspeedresearch

70% of Americans say they consult product reviews or consumer ratings before making a purchase. psbresearch

80% of consumers prefer to shop online because it is easier to research products. e-tailing

User reviews are not enough!

Current solutions focus on customer reviews. This approach suffers from an insurmountable fallacy. Products without reviews are less likely to be bought; therefore, the chance of someone writing a review is smaller. Unless a web store has enough customers taking the risk to buy a product with no reviews, it will have a hard time getting reviews for more than 22 percent of their products. This puts small and medium stores at a big disadvantage!

Increase your sales with expert reviews on your site!

Dekiba provides online shops with reviews for eighty percent of their products per vertical. We make sure that 80% of your products have reviews right there on your site. Our review visualizations are based on multiple expert reviews for every product.

The integration of Dekiba reviews is very easy.

Dekiba is currently in beta.

We started with a few selected product types and we are expanding our product catalogue according to demand.

Dekiba’s technology is based on a combination of algorithms for text categorization and opinion mining that produces many more detailed and accurate results than any solution available so far.

Dekiba analyzes expert reviews and aggregates highly accurate and relevant knowledge about every feature of a product. Using statistical models to interpret and a combine this information, Dekiba can offer very objective ratings and product intelligence.

Our algorithms for text categorizations are inspired by recent advances in neurological research. Utilizing this approach boosted the recognition rates drastically. It takes humans several days of training to gain the same accuracy as Dekiba’s current algorithms.

Most existing solutions use standard binary sentiment classification while we use hierarchical and multi-dimensional sentiment classification. This approach caused the initial development to be much more demanding – but our high recognition rates would not have been possible otherwise.

Dekiba uses machine learning approaches, such as genetic programing and logistic regressions, to optimize self-learning algorithms. This helps us continually improve the scalability, performance, and recognition rates of our text classifiers.

Dekiba started as two high school friends who decided to turn their passion into a business. Both were guys whom friends and family consulted whenever they needed to buy electronics. They followed new product releases because they were passionate about the topic.After high school, Arnold Keller studied Computational Biology in Austria. In the course of his MBA, Emanuel Schattauer moved to Japan, specializing in Neuroeconomics.

It all started with a simple problem. During his stay in Japan, Emanuel needed a new laptop. Being proficient in German, English, and Japanese allowed him to do product research in all 3 languages. To make the research more efficient, he developed a system to quantify the results of reviews so that the text’s language didn’t matter. It became obvious: the information gathered would be very useful to everyone in the market looking for a new laptop.

On a visit to Austria, Emanuel convinced Arnold to drop out of the university and join his startup as a co-founder. The combined skills of the two entrepreneurs brought insights that others in the field were missing.
Arnold’s experience with sequencing DNA proved to be invaluable for analyzing text. Many natural language processing and machine learning techniques are employed for computational biology.
Emanuel was captivated by Neuroeconomics. His thesis explored the macroeconomic effects of algorithms employed by the human brain for calculating value over time. His continued interest in the field paid off; several of Dekiba’s most-successful algorithms were inspired by advances in neurological research.

Dekiba is now a team of five people working from different places around the globe: Japan, the Philippines, Europe, and the United States.

Obviously, our academic backgrounds had a big impact on us. But being immersed with technology since our youth was another strong factor. One rather surprising influence came from us being avid gamers. To illustrate this point, we accumulated a huge database full of dense information. But how could one make all this data intuitively understandable to everyone?
We found out that this problem had already been solved. Visualizing complex information was an essential task for game designers. As an example, game designers had to inform the player about how much damage a weapon caused, how fast it could hit, what magic effect it possessed, and so on.
We decided to build our system on decades of research and optimization. Gaming is a multi-billion dollar industry which taught us so much about how to engage users. For the first time, e-commerce can take full advantage of this data and technology. We believe the online shopping experience is waiting to be revolutionized.

• We want to improve people’s lives. One way to do so is by helping them to make better and easier purchase decisions. Additionally, the demand of the market guides us to create something people actually want.
• Another motivation is to level the playing field for e-commerce businesses. Independent of size, all online shops should be able to provide shoppers with the best information available.
• Furthermore, we want to contribute to innovation by helping manufacturers focus on creating better products.

Today, Dekiba’s technology is excellent at extracting knowledge from reviews. We estimate that our current algorithms are at the level of a 12-year-old. Our goal for the future is to raise our comprehension level to that of an adult. Ultimately, our approach will provide the basis for an artificial intelligence which can understand complex concepts found in news stories or literature.