An Interview with the Co-Founders of Genba

What are some of the macro challenges faced today by technicians in the manufacturing landscape?  

"Let me first setup the landscape then dive into the challenges…

Within the industry there is a group of highly skilled workers called technicians. They have the hard job of keeping the plant and assets up and running by responding to planned and unplanned downtime events. The knowledge to keep the plant running is mostly in their head. There is a huge transfer of knowledge between techs across shifts to get these plants to run well.

Much of the knowledge walks away - literally the people involved in the process don’t log the information before walking off the plant floor for the day.

Further, there is a generational gap occurring in manufacturing today with senior technicians retiring at a faster rate than skilled workers entering the labor pool. Talent is walking out the door with over 20-years of knowledge in their head. Layoffs and early retirement during covid has only exacerbated this challenge.

Existing software doesn’t work well because it’s primary use case is to track assets, time, and inventory. This is not designed to facilitate comms and knowledge transfer between workers.

Today there are a number of work arounds: books, sticky notes inside panels - aka, caveman drawings - whiteboards, and open-text fields inside CMMS (computerized maintenance management systems). There’s truly no good system of record for knowledge management, so you get all these disconnected systems.

This knowledge breakdown has tangible cost associated with a lack of visibility into the nuance happening on the floor across shifts and across roles in operations, safety, managers, and reliability technicians.

This knowledge transfer problem is at the core of what Genba is solving."

"There are additional facts which are compounding this challenges faced by technicians.

Maintenance is inherently a knowledge skill. Whenever assets are installed into the facility someone has to maintain them. You can only google troubleshooting steps for so long - or ask others - after that the issues become highly contextual to that machine.

It’s all very tribal - who maintained, how frequently, last PM v. unplanned downtime event, etc. It takes trained humans to see the connections between these events."  

What anecdotes & data are you bringing into your research when solving this knowledge challenge?

"At its core, maintenance is a cost center. Manufacturers are squeezed to do a lot with very little and talent is no different.

Knowledge transfer is so important in these lean manufacturing environments. More important is getting this knowledge to flow in multiple directions.

I was just in a plant last month and found out the machine operator was frying chicken 2-weeks ago - now he’s maintaining a complex machine critical to high-value assets inside a plant. So while it’s fantastic to see real people getting started with lucrative careers, this skill-matching poses a tremendous challenge to manufacturers."

"Design is central because humans work in industrial manufacturing!"

"Up-skilling & re-skilling is key.

I was recently on a McKinsey webinar on the state of Manufacturing. During the presentation there was a poll to the audience of mostly General Managers and VP-level industrial manufacturing executives at high-value asset producing plants.

What is the biggest challenge in supporting the strength of US Manufacturing. The MOST VOTES went to 'talent investment'. Even above technology and capital investment! So those running the plant at the highest levels are figuring our where to invest to bridge this knowledge gap."

How does technology aid in these up-skilling & knowledge transfer charters from the financial perspective of the maintenance technicians within large, high-value asset manufacturers?

"The primary job of maintenance technicians is to maintain assets. Don’t maintain more than the machine really needs but don’t under-maintain, either. Both of these situations lead to increased cost in the form of inventory management and worse, unplanned downtime from under under-maintenance. The use of resources as efficiently as possible in order to run the facility as efficiently as planned is a tough balance.

I’ll give a very real example - lack of knowledge in a shift-changeover about a specific assets causes an unplanned downtime event. This downtime stops a really important shipment from going out the door. Further, if something is not maintained properly between shifts, the longevity of an asset is compromised. This relates directly to top line revenue and machine expense."

"There’s a huge desire to ensure the workforce is trained to use modern tech to their advantage and to maintain an ever complex plant. More automation goes into the plant every year. This takes away from operations staff as the assets are more difficult to maintain. There is also a growing requirement for highly skilled technicians with college degrees."

"This is compounded by 3 things: a talent squeeze at bottom, an exodus of talent at top, and especially in 2020, a reduction of capital investment."

Where is Genba unlocking value in the process of asset management?

"I’ll start by addressing maintenance work. This work falls into two categories.

Planned downtime is just as it sounds. The asset is scheduled to be down for a period of time and work is planned so not as to impact high-value asset production. This occurs at set intervals based on an assets life cycle or run time.

The flip side is unplanned downtime. This downtime is tremendously costly to the organization and by its very nature, can be challenging to find the root cause. In order to solve the problem, technicians need to be able to quantify various forms of contextual information. Once complete, that same human needs to go into the CMMS and input the myriad of fields to recreate a problem, often hours or sometimes even days after the job. Companies create drop down fields for the 70+ potential root causes - this is always expanding - and the technicians need to carefully select which one was correct.  

Further, this information needs to be codified by managers in order to analyze and guess when it might happen again. At the end of the month another human needs to go in and use Pareto rule to find a cause. This manual, inefficient way to extract information from the CMMS is slow, untimely, and often inaccurate based on the inputs and analysis lacking context."

'There is software which offers the nouns, and there is software which understands the verbs. Genba meets technicians where they are, and understands their context and intent...the verbs.'

"When the asset goes down, a technician’s mental model is to build context of what happened. A lot may have contributed to the asset going down and several question get asked:

Who touched last? Has the operations team seen this shift and last shift? What iOT data is available? How often is this going down?

So the context is required and incredibly important."

"Right. Today, extracting context is almost impossible. In some systems you can’t even search comments in past work orders. People are clicking one-by-one into dozens of work orders to answer basic questions. It’s very common for techs to spend over an hour to build this mental model and piece everything together."

What about Genba’s technology is helping to piece together contextual information?

"Technology like Speech-to-Text (STT) and Natural Language Processing (NLP) are maturing to unlock the core job. NLP to understand intent and find patterns is just now available to market... yet it lacks industry-specific context.

Our combination of specifically trained speech models allows Genba to provide context to the STT, saving technicians time in the process of work order knowledge capture, while also providing much more accurate transcriptions.

The technician is going to get more powerful and effective when you give them the ability to rapidly compare qualitative + quantitative + sensor / vibration data."

"There is a heavy trend to use iOT to do predictive analytics. The technology has made advancements yet it’s not the only data that is out there. What’s not easy to quantify is unstructured data like troubleshooting notes and work order history. There are tons of insights in that data, but not much is being done with it right now. From a technician’s point of view, trying to digest loads of IOT data, logs, and graphs during a downtime event is inefficient.

Genba is able to use the context of the situation with NLP & STT to provide data in new ways at all levels of the manufacturing organization.

In the future, Genba will be able to go back into the asset history and give recommendations in real-time. In this way, Genba is a knowledge transfer and learning tool for re-skilling and up-skilling workers."

"There’s a lot wrapped up in the problem and it’s less of a jumble of problems, and more of a hierarchy. Genba starts with the core workflow for a single shift, then enables workflow from multiple shifts, and eventually multiple plants. It's not just technicians that will benefit from Genba--operations staff, reliability engineers, maintenance managers, plant managers--everyone wants clear visibility into what's happening on the floor and how it's impacting core KPIs."

What role does design play in solving this problem?

"Design is central because humans work in industrial manufacturing!"

"I’ll elaborate, design is critical in an industrial manufacturing setting, because those settings are loud, dirty and stressful, and workers are not going to fumble around with poorly designed technology. It’s even more important than within consumer technology. Good design and reliable software is paramount."

"This doesn’t just mean it looks pretty. That’s important, but it has to be functional. Our CTO has a nice analogy. 'There is software which offers the nouns, and there is software which understands the verbs. Genba meets technicians where they are, and understands their context and intent... the verbs'."