Why product metaphor?

Wronged by Excel.

Once, the finance director called me and asked me to train her analysts and accountants in advanced Excel functions. I talked to the chief analyst and accountant, who detailed the request. They had a problem with the order fulfillment spreadsheets the suppliers reported to us weekly. The report format was unclear, and there were no controls, so analysts got garbage at their input. And they wanted Excel to find incorrect data, change it to the proper format, make a list of corrections the provider needed to fix, and send him an e-mail with these proposed corrections. ‘You got it all wrong,’ — I said to them. ‘The first thing you need to do is figure out what report format you want, and for that, you need to define what records structure you need to do your job. Now you don’t have either. Anything you would do with Excel checks and workflows will be useless.’ But I couldn’t deliver my message; they wanted Excel training for advanced users to fix the wrong problem.

I’ve been thinking about this case for a while until I figured out the root cause. For me, after I read a beautiful book (Sawyer, 2009), Excel, for many years, was a programming and modeling language that required data structures and algorithms. I knew each row was a record with a type, and one table was a collection of records of one kind. We define the class with attributes in columns containing types (integer, date, money). If I wanted to build a report and select some records in the table, I first needed to understand what class I wanted to declare. I had an evident and rich product metaphor (that was a ‘simple database with graphics’). It was not the same for the trainees. For them, it was a lined piece of electronic paper with a built-in calculator — no types, no data structures, and no algorithms to transform structured inputs into structured outputs. The same product for us seemingly kept a completely different meaning and value because we have had other product metaphors.

A cart or a car?

Another time we were doing an online store logistics optimization project. We have had two potential bottlenecks — one in the last-mile delivery and the other in the fulfillment center. I was responsible mainly for the latter. We implemented all our ideas, but it was not enough; we still required at least a 20% improvement in the number of orders shipped during the day. But we ran out of ideas and had to get out of the box. During one of the meetings late in the night, when we were going through the warehouse concept model, I questioned the basis of it, the warehouse metaphor.

My idea was simple; currently, we organized the layout and workflow around order execution. One operator got one order and picked it up all positions in the order. We assessed operators’ performance based on their speed. So operators hated orders that required picking parts all over the warehouse, which required a lot of movement and search, so their performance and income suffered. And we have had a lottery feature in the warehouse software that accidentally distributed orders across operators. But you know how people think and find patterns in pure coincidence. Operators believed that the game was rigged and that there was bad luck. The warehouse layout and principles didn’t work well with B2C orders, and that caused the bottleneck. That was my assumption.

I came from the automotive industry, so the first idea that came to my mind was to change the warehouse concept to the production line approach. In this line of thinking, you would not consider the fulfillment center layout as a warehouse; it is a shop floor, a store. So first, we rearranged storage space based on orders’ positions and proximity numbers. Cart must travel as short as possible. Second, we allocated operators on shelves, which became our processing centers. And finally, we changed KPIs to assess processing centers and overall throughput. The simulation of the proposal worked perfectly, but operations considered the transition project too risky and went through only during the COVID lockdown, when demand surged several times, and there was no other solution. Again, the change in the product metaphor reframed the product completely.

Product metaphor market fit.

Another story I already wrote is the one about Reed. He saw the ship as a formula, the final metaphor, the most abstract one possible to imagine. Others didn’t see it and their products failed terribly.

Metaphor is the key to getting the idea of a product — user insight into the product is as essential for product community growth as are patterns of use, the cookie cutters. Without a clear understanding of metaphor, one cannot make product decisions that will result in positive change in the emergency of valuable and usable products because the users will never understand the product concept. And for me, this is a question that language-market fit experiments (Lerner, 2019) answer. How well do you know the product metaphor (have you completed solution discovery?), and how can you communicate it to the target audience?

But if metaphors are essential, where and how do we create them? What are the quality criteria? Also, for product growth, it is critical to know whether users and customers have the metaphor in their mind and if you can elicit it in a marketing message or need to educate the market.

We inscribe the metaphor in the product design, the way Don Norman (Norman, 2013) taught us. An ideal analogy would be the Casio lighting keyboard synthesizer. Just hit the red buttons, and you get a melody you like.

Bibliography

Lerner, M., 2019. Finding Language/Market Fit: How to Make Customers Feel Like You’ve Read Their Minds. URL https://review.firstround.com/finding-language-market-fit-how-to-make-customers-feel-like-youve-read-their-minds (accessed 4.17.22).

Norman, D., 2013. The Design of Everyday Things: Revised and Expanded Edition, Revised edition. ed. Basic Books, New York, New York.

Sawyer, T.Y., 2009. Pro Excel Financial Modeling. Apress, Berkeley, CA. https://doi.org/10.1007/978-1-4302-1899-9