Come on a virtual tour with us – every Thursday at 2pm!

How matching works

you’ll experience this first hand if you select the participant role

What are the functionalities

a behind the scenes look at the backend for organisers

What platforms can do for you

find out via case studies from our CHEMISTREE clients

Intelligent Matching

Connecting people in a targeted manner is an art. Depending on the scope and complexity, it can also be a mammoth task: Collecting, comparing, allocating relevant information… and as soon as anything changes somewhere, rearranging it again. With digital matching, the unlimited power of IT comes to your aid. Optimal connections between participants are determined on the basis of digitally collected data. Individually tailored, according to a standardised and automated procedure and as often / as quickly as you want. CHEMISTREE’s algorithm-based matching process promises you technically excellent, scientifically sound and ethically responsible matching – tried and tested thousands of times and constantly optimised.

The CHEMISTREE Matching Formula

Optimally Combined Algorithms

We use Nobel Prize-winning algorithms for scoring and matching and don’t combine them according to pre-set formulas but tailor them to your project. If you like, we can let you go down to the smallest parameters. We believe, there is no better way to pre-program trust.

Well-founded Questionnaires

Expertise, role, needs, approach, … no matching questionnaire is like the other. We work with you to design the content and logic and tailor it to your project. An integral part: The set of personality preference queries developed together with the LMU as a foundation for productive, stable tandems

The CHEMISTREE Matching Formula

Specifically Designed Fit

We define the ‘perfect fit’ as constructive for your project and leading to results: similarity and/or diversity, hard and/or soft factors, … all factors are specifically related and weighted. No ‘one size fits all’, but tailor-made matches for your goal!

Humanly Harmonious Tandems

Even when the technical and content-related match is overwhelmingly good: if there is no personal chemistry, tandems remain well below their potential. CHEMISTREE works with psychologically sound questions on personality and creates that elusive chemistry which is needed so people can inspire and motivate each other and optimally perform together.

A winning formula for tandems

CHEMISTREE carried out a joint study with the chair for psychological methodology and diagnostics at the LMUniversity Munich on 400 test subjects in 2018. We wanted to know: What is the most important lever for personality-based matching? The surprising finding: Tandem partners with very similar personalities are more satisfied when working together – but not more successful! They are surpassed by tandems in which the preferred personality traits of the other person apply. Exactly 11 characteristics were determined in the study that lead to significantly higher goal achievement in tandems. As a psychological questionnaire, they are an important part of our matching.

The two renowned psychology experts Prof. Dr. Markus Buhner and Prof. Dr. Felix Schönbrodt carried out a matching study in 2018 together with CHEMISTREE. They describe the scientific background of matching and the formula for success for tandems, which is in the CHEMISTREE matching process:

Fit is more than filtering…

Whether thematic, hierarchical or personal: Sometimes like likes like – and sometimes diversity is important. A real fit is complex: considers ‘hard facts’ and ‘soft facts’, weights the criteria in relation to each other and combines the same and/or different – depending on what the objective is!

… and only good data result in good matches.

With matching, the quality of the result depends on the quality of the data on which the matching is based. That’s why CHEMISTREE is by your side when designing each matching questionnaire: With our experience from a wide range of customer projects, we develop the optimal set of questions together.

You can save yourself the brain teaser!

It takes a whopping 495,000 arithmetic operations to form all the tandems that are possible in principle from 100 people using 100 differently weighted criteria. In order to ultimately achieve optimal matching and to make sure no one goes empty-handed, 99 double faculty variants have to be played out. That’s 27 253 921 397 507 295 029 807 132 454 009 186 332 907 963 305 45 803 413 734 328 823 443 106 201 171 875 possibilites!

Manual Matching as Brain Teaser

Algorithms – Friend or Foe?

Algorithms are not inherently ‘good’ or ‘bad’. They represent decision-making rules – all weaknesses and errors that are part of human decisions are also reflected in an algorithm based on these. That is why different perspectives are needed as early as in the development phase to ensure distortions are not programmed in in the first place. You also need transparency to be able to check applications in their rules for possible imbalances. Ultimately, intervention options are required in order to be able to correct critical parameters. All of this makes the difference between ‘good’ and ‘bad’ algorithm-based software and AI.

As a provider of algorithm-based processes, we feel we have both a technological and moral responsibility: from the development to the conception of an individual project, our matching process should be objective, distortion-free, transparent and self-determined by the participants. Furthermore: Not only WE want to meet the requirements of an ethically responsible AI – in our opinion, ALL algorithm-based systems that affect people in any way should do it. To this end, CHEMISTREE participates in federal government initiatives at various levels.

The KIDD process

Aims for a fair, transparent and understandable application of algorithm-based systems and AI in companies by involving a variety of employees and experts during their development.

The AI Standardisation Roadmap

Defines requirements and challenges as well as standardisation needs for seven key topics of artificial intelligence.