Sourdough & Lean

A few fails along the way but finally a success after a week of cultivating sourdough starter.

It seems that during these weeks of Covid-19 lockdown, people fall into 2 camps. Those who make sourdough and those who don’t. I used to be in the latter camp and didn’t understand what all the fuss was about, but after a week investing in a smelly starter on the bench, I’m now officially a sourdough baker.

What has sourdough got to do with lean? Well, it’s just a process. Lean is all about looking at processes, continuous improvement and the engagement of people. Judging by my first attempts at making sourdough, improvement was definitely required.

Successful organisations are those with a continuous improvement culture. But organisations are just made up of individuals. When managing change, we can look big picture (vision, skills requirement, incentives, resources, action plan etc.) and we can look at where each individual is at in the change process (awareness, desire, knowledge, ability or reinforcement). It’s the same with continuous improvement. In the case of me making sourdough, there is no big picture. I don’t have to get better at making bread. There are lots of great bakeries in South Melbourne. There’s no burning platform. For some reason I was invested in the process of making sourdough enough to want to get better at it. So why was that?

I was engaged in the process, enjoying learning something new. I realise it’s only worth talking to other sourdough enthusiasts about it. Otherwise I can see their eyes glaze over (even over Zoom). If you’re not engaged, you’re not interested in improving. This applies to any process.

At a recent client, one of the team leaders regularly laughed at me when I got excited by the possibilities for improvement in some of her team’s processes. She didn’t get it but was happy to humour me during coaching sessions. A few months later, she came up with a brilliant idea to set up assembly stations for products, drastically reducing waste (in particular overproduction, waiting, motion & inventory) in the process. She was genuinely excited by the changes and the improvements and kept building on her successes.

The best ideas for improvement come from people within a process. They know the detail. They might need help with how to fix the problems, but they usually know what the problems are. I know a lot more about making sourdough by getting my hands dirty than I could ever learn from YouTube.

Control the inputs, not the output
Although organisations measure the outputs – safety stats, financial performance etc., these cannot be directly controlled in the same way the inputs can. Outputs are used to feedback whether the input control needs improving.

In making sourdough, the experts all recommended weighing the flour, feeding starter at the same time, keep the temperature the same ie control what you can. But as sourdough starter has life of its own and is variable, some inputs could not be controlled enough to guarantee a good result. This is where control points along the way are needed. Adjustments to time or temperature depending on how much the dough has risen etc. I learnt what to control and what to measure along the way to increase the chance of an Instagram-worthy product.

It was exciting to see how the bread was going. Feedback was the sight of a great looking loaf coming out of the oven and in the tasting. Similarly as in the above example if the improvements can’t be measured there’s no real incentive to keep doing it. It’s why teams keep score, or you try to beat your PBs. We like feedback to see how we are doing.

In continuous improvement speak, we call that process PDCA, or Plan, Do, Check & Act. By checking results and making adjustments that feed into your next plan (or next batch of sourdough) you keep the feedback loop going and keep getting better.

So, making sourdough is just like any process. Control the inputs, measure the outputs and feedback results. Engage those within the process and get improvement ideas from them.

Making sourdough is better than many processes I’m usually improving though, as I get to eat the output.

My Mini

A few months ago, I got a new Mini. Very exciting as my old car was almost 20 years old. Minis are made to order and shipped from England, so it was a 3 month wait. Worth it. But after a couple of weeks, the back bumper looked crooked and rubbed on the tailgate, causing the bumper to get scratched. I took it back to sales, who asked me to speak to warranty. Warranty passed me onto service. Mini service was being renovated, so I had to go to BMW service a couple of km away. Service took it in, gave me a BMW hire car and passed it onto their body shop. I rang service every week to see how my car was getting on and eventually 5 weeks later I was told that it had been fixed.

I went in to pick up my Mini and the bumper looked the same. Still crooked, but the scratch marks had disappeared. I pointed it out to the service man, who called over the warranty woman. They both looked at it and agreed that it had not been fixed. So I left in the hire BMW and my Mini went back to the body shop for another go. Another week later, I picked up my car. All fixed. But a few days later it was crooked again.

This time I went straight to the body shop. They fixed it on the spot. But after a few weeks later the bumper was crooked again and getting scratched again. I took it back (again). They booked it in for another visit to the body shop, another respray and I had another hire car for 2 days. This time, however, they appeared to get to the root cause of the problem – a faulty bumper clip – which they replaced. The bumper is now straight and doesn’t rub on the tailgate.

As a customer I had to deal with 4 departments: sales, warranty, service and body shop. They all had different priorities on cost, time and quality. Body shop told me they had a spare bumper and wanted to replace it, but warranty told them to fix the original. Service was telling body shop it was urgent, but there was no follow up to get it back to me. I just wanted my car fixed within a reasonable timeframe.

Some lean concepts or tools highlighted by this:

Value Stream Mapping – if Mini looked at the number of people that had to interact with the customer, they might reconsider their process.

Misaligned KPIs – warranty, body shop and service all appeared to have had different things being important to them.

Right First Time – obviously if the bumper had not been faulty at the beginning, there would have been huge cost savings for Mini/BMW.

Communication – So important. Communication both between departments and with the customer could have been better.

Waste – DOWNTIME – Defects, Overproduction, Waiting, Not utilising talent, Transport, Inventory, Motion, Excessive processing – at least 5 wastes identified here.

The whole experience did make me wonder whether Minis and BMWs were expensive because they were great cars and therefore provided value for money or whether they needed to cover the wastes in their processes.

I still love my Mini, though.


Keeping Score – KPIs

Why do we have KPIs (Key Performance Indicators)?

If you’re playing sport, you usually keep score. Scores tell you whether or not you are improving. Footy players keep score. Athletes keep track of their PBs (personal best). Coaches measure yet more statistics. KPIs are just the scores or measures for your business or organisation.

Most organisations either have too few measures, or too many. I recently worked with a business which supplied products to the building industry. They all agreed that the reason their customers chose them was for the quality of their products. Yet they had no quality measure anywhere in their business.  It would be very hard for them to know how they were performing and whether they were getting better or worse. They were relying on their own installers to fix up quality problems before the customer became aware of them, but at what cost?

It’s important to get the balance of measures just right – both in the number of measures and in the spread or balance. At organisation level, this is often called a balanced scorecard. Not everyone in the organisation needs to know and manage all the measures (just like the players don’t know the level of detail the coaches do on a footy team) but the measures any individual has needs to mean something and be within their control.

The KPIs for an individual or a department must also drive the right behaviour and be aligned with the organisation’s targets.  Be particularly careful of department specific measures – those ones that make a department look good at the expense of others. An example is purchasing having raw material unit price as a measure. Reducing unit price can drive ordering in large quantities, causing headaches for the warehouse, increased storage costs and risk of damage or obsolescence. It might also result in decisions being made on price at the expense of quality. In this case the purchasing department also needs to have overall stock value or stock turns and product quality or warranty costs as a measure.

Keep it simple. Just as players are too busy playing the game to study the stats, operators are usually too busy working to study pages of numbers, even if they are visually displayed. 4-5 measures is a manageable number. Number goals (e.g. good parts produced) often means more than percentages. The measures need to be within the person or team’s control.

Keep it balanced. People will do what is asked of them, as they will think that is what is important to the business. So if production volume is the only thing measured then quality or safety may suffer.

Priority of KPIs or measures

Prioritise KPIs in the following order, so that people get to learn what the organisation’s priorities are. Quality is high up the list because in most cases, if you try to drive productivity, quality will suffer. However, if you focus on quality and improving processes, productivity will improve:

  1. Safety
  2. Quality
  3. Delivery
  4. Productivity
  5. Improvements – projects or ideas

These points can form the basis of a tool box talk agenda, or month end review meeting.

Some examples of possible KPIs or measures are below. But make sure you review your measures and targets regularly to ensure they are driving the right behaviours.

Suggested KPIs or measures

  1. Safety – no of days since MTI/LTI (medically treated/lost time injury). Don’t have a target on first aid injuries, as people may be tempted to hide minor injuries. A great proactive measure is the number of times someone is ‘caught’ behaving in a safe manner such as wearing the right PPE. 5S (workplace organisation) score or no. of 5S audits completed is another option.
  2. Quality – no. of parts made right first time (RFT). This should be recorded positively and is one example where percentages works OK as % RFT of total parts (similar to % free throws by a basketball player).
  3. Delivery – DIFOT (delivery in full on time). The main problem with this is many sites change the dates – usually with the customer’s agreement – so always have 100%, which doesn’t reflect the chaotic reality. Schedule adherence for manufacturing is good as it is more within their control and is a good indicator of stability.
  4. Productivity – no of items per team or cell per hour or day. Or number of batches or jobs. Ideally there is also a per person or per manhour measure so that if one if the team is missing, the rest pull together to make their numbers look good.
  5. Improvements – no of projects on the go, no of projects completed, savings as a result of these projects, training hours. This may be an input measure. A footy team might measure attendance at skills training sessions.

Input vs Output measures

This is a blog post in itself but if you trust the process, getting the right inputs will result in the outputs you want. Checking outputs is often too late.