Data-Driven Design

Data-Driven Product Design

A typical product design process is a knowledge-driven process with very little data involved. Product designers and engineers tend to trust their knowledge and expertise as the main resource for product design decision-making and product brief writing. Involving data in a product design process can improve the design quality, reduce costs and shorten time-to-market. In this article, I’ll explain what a data-driven product design is, what its benefits are, why it’s not common, and how to implement this useful method.

What is the difference between data and knowledge in the context of product design?
Data is a collection of individual facts. For instance: a list of materials’ prices. Knowledge is a meaningful acquaintance with facts, principles, methods, or practices to understand or perform a specific subject. For instance: the ability to choose the right materials for a product.

Data without knowledge is meaningless. The data is (among other things) the knowledge’s structure. The knowledge is the meaningful needed skill to make decisions and act.

Knowledge is more complex than data. Yet, the human brain can comprehend and remember knowledge much better than data because knowledge is based on logic and data is based on many random details. This is why the human brain prefers a knowledge-driven work over a data-driven work. It’s simply easier and faster to perform.

What is the problem with a knowledge-driven design process?
There is no problem with a knowledge-driven design process (in fact, it’s a must), but rather with the lack of a data-driven design process. Knowledge is based on data, and some data tends to change over time (suppliers, prices, new materials, etc.). Continuously using the same knowledge without refreshing it with updated data can lead to inaccurate knowledge.

Let’s see how a product’s materials-choosing process would look with a knowledge-driven process and a data-driven process. Knowledge-driven process: The engineer chooses the product’s materials based on his knowledge and then reviews only the chosen materials’ data (sheet) for final approval. Data-driven process: The engineer searches and reviews a wide range of optional materials’ data (sheet) and then chooses the best materials based on the data.

The knowledge-driven process is faster, but it might result in missed opportunities for better and more proper materials. The data-driven process requires more time and tedious work, but it can lead to a better and smarter decision making process. A data-driven process is not replacing a knowledge-driven process, but take place alongside with it.

Product designers and engineers tend to skip data-driven design methods because it requires a lot of data searching, analysis, and because there are no good tools to support this process.

Implementing a data-driven design method
The main problem with a data-driven process is the huge amount of data involved in any of the product’s attributes; you just can’t find and analyse it all. I recommend choosing one or two main attribute(s) of the product—the one(s) with the most potential to improve the product, and then to perform a data-driven process only on for these attribute(s).

Implementing a data-driven product design method alongside the common knowledge-driven design method can improve the product design quality, stimulate innovation, and reduce time-to-market. Product designers and engineers should make the needed effort to learn how to use this this valuable tool and implement it.


My next post will be about “Designers’ tools for non-designers
Be notified about this upcoming post.

  • Christian Theill

    Yariv, this time I disagree with your thoughts. I can’t share your distinction between knowledge and data driven design. Here is why:

    If we consider knowledge only what we once learnt or experienced you are right (and, alas, the world is full of that people). This kind of knowledge in fact may be based on surpassed data.

    If we consider knowledge as a dynamic process, updating (up-to-date means also up-to-current-data) it daily and especially when facing new projects, the concept itself of knowledge includes the usage of the data’s state of art, and there is no need of distinction between both.

    This is what real knowledge means to me, not only in design but in every field.

    What you call data driven design isn’t anything else than updated knowledge driven design. And this means that we can keep trying to be better than others remembering every day that our knowledge is the one of yesterday.

    • Yariv Sade

      Hi Christian, thanks for your comments.

      The known DIKW pyramid (or hierarchy) structures the relationships between Data, Information, Knowledge, and Wisdom ( As I see it, there is no sharp border between each of the four types of knowledge, yet there are some significant differences. Data is a (normally large) collection of facts without meaning. Information is a higher level that give some meaning to the data – it answers who, what, when, and where. Knowledge is “above” information and it answers mainly “how” questions. Wisdom is at the top of the pyramid, and it focuses on “why”. This is, of course, a schematic structure, in real life it’s sometimes harder to distinguish between these four levels.

      You are claiming that knowledge must be built on updated data, otherwise the knowledge meaningless. You are right, but most people (not all), once they have reached the knowledge stage, they never go back to the data-lake, to search, refresh and update their knowledge with updated data. They don’t search for and analyse newer data, but just confirm their knowledge decisions with very narrow needed data. That’s the whole idea behind my post.

      • Christian Theill

        Yariv, you confirm what I said: if knowledge is intended as a STATIC state of assumption and elaboration of data and information,then it is not real knowledge. The DIKW pyramid is a good scheme but it can’t have a static read since its D-factor is continuously growing. I, K and W lose their sense and can’t be defined as such if they don’t follow the same DYNAMIC growth.

        Also before your appreciable approach there was a term to define those who stuck ages ago with their “knowledge”: we used to call them PRESUMPTUOUS.

        Strange enough that most companies bet on the (static) scholastic knowledge of their engineers, thinkers and developers not demanding its dynamic, evolving approach.

        I rarely aspect myself the W of the pyramid, wisdom, but I often observed that awareness of real (dynamic) knowledge leads to it.

        • Yariv Sade

          Christian, I believe we are saying the same things. While I was describing the actual (common) situation, your are describing the expected situation.

  • Jos Voskuil

    Yariv / Christian – I see two sides of the story. I support implementations of PLM in many companies and when it comes to knowledge companies are trying to capture and update their knowledge base built by experiences from different engineers described in documents.

    In some companies even searching and browsing through this knowledge base is to much for some engineers who have all the knowledge in their head, therefore missing opportunities to improve and adapt. It is human behavior.
    One of the underlying reasons for this behavior is that most of the time having and protecting this knowledge is the job security for this engineer and most of the time they did not grow up with easy accessible data through internet.
    If am sure future generations and future processes will be more and more data-driven to provide better knowledge. This is what Christian is stating and I agree.
    Reality is that in many companies data is still outside

    • Yariv Sade

      Jos, thanks for sharing. I agree with you. I sure believe that data will be more connected to knowledge in the near future.