If you’re struggling to grow your business, you only really have two options. One is to systematically find the root causes of your limits to growth and fix them. The other is to focus on making the symptoms of your growth problems go away, fall into a cycle of stagnation, and wait for the board to lose patience with you.
This happens all the time. Companies invest heavily in growth initiatives without first understanding where their real bottlenecks are. It’s like taking aspirin for a headache that comes back every day. You’re treating the symptom without addressing the underlying cause. And the consequences for mid-market companies can be existential.
In this article, I’ll give you a straightforward approach to finding and fixing the constraints that are actually limiting your growth, based on a framework called the Theory of Constraints. It’s not particularly complex, but it requires discipline and a willingness to look beyond the obvious.
Symptoms vs. causes and why most companies solve the wrong problems
The temptation is always to solve the thing that hurts the most. Unfortunately, that’s usually a symptom, not a cause.
Let’s take a concrete example. Imagine you have a manufacturing line with two bottlenecks. It’s a linear process, and you have a bottleneck at step two and another at step five. Step five is right before you ship to the customer, so it’s the one that’s really visible to everybody.
Under pressure to improve output, you decide to solve for step five and invest in a second shrink wrap machine (the visible bottleneck). But after implementing the “fix,” no more products are shipped than before.
Why? Because you solved the visible bottleneck without addressing the upstream constraint that was actually limiting throughput. You doubled capacity at the end of the line, but you didn’t need that capacity yet because you’ve got this earlier bottleneck that’s not been solved.
Let’s make this more concrete. What if the earlier bottleneck is due to a quality check step that finds lots of defects and has to send products back for rework? And what if the root cause of this problem is an aging manufacturing component that’s beyond end of life, making every device that gets made a little bit worse than the last?
The reject rate gets higher and higher, and suddenly your expensive new wrapping machine is sitting idle because you’re not getting enough product through the line to keep it busy. Until you fix the machine that’s producing defective products, you don’t really know how many wrapping machines you need or what your true production capacity should be.
While my example is from manufacturing, this pattern repeats across every kind of business. And it has cascading effects:
- You can’t make your sales numbers if you can’t deliver the product
- You can’t meet customer expectations if your product or service is riddled with defects
- You can’t scale your marketing funnel if your conversion process is broken
The theory of constraints: A framework that actually works
There’s a practice that I really like based on the Theory of Constraints, or ToC. It was originally put forth by Dr. Eliot Goldratt in a corporate fiction book called “The Goal,” and more recently adapted to IT in another book called “The Phoenix Project” by Gene Kim.
To put it very simply: find the bottlenecks, understand their effect on the process, and then either eliminate them or manage them. That’s basically it.
The power of this approach is that it forces you to understand your business as a series of interconnected processes. And let’s be honest, 80% of everything that happens in your company is cyclical. It’s turning the crank: whether it’s qualifying leads, getting people to fill out an application, collecting documentation, scheduling closing appointments – we do these things over and over.
The question is: can you produce the same level of quality output every time you turn the crank? Can you do it reliably? Can you do it better? Can you scale it economically? And most importantly, where are the constraints in your system that are preventing you from doing this?
Four types of bottleneck that kill growth
In my work with companies of all sizes, I’ve found that most growth-limiting bottlenecks fall into four categories. Each one requires a different approach to identify and resolve.
Capacity bottlenecks
These are the most straightforward: you simply don’t have enough resources to process the flow of whatever you’re moving through your system.
Here’s a tech example: let’s say you’ve got a very large development organization, and all of your software builds and deployments are orchestrated on a tool called Jenkins. Jenkins is great, but it doesn’t really scale well. If you have hundreds of projects flowing through it, you can scale up, but you have to invest in lots of resources to do that.
A lot of companies start with a single Jenkins server and pile on more projects as they grow. More builds, more complexity, more developers. And then one day all the builds are failing because Jenkins is overloaded and crashing. It has become a bottleneck.
The same principle applies to call centers, customer service teams, loan processing departments, or any resource that has a finite capacity. The key is identifying the capacity constraints before they become critical and either expanding capacity or finding ways to manage the load.
Quality bottlenecks
Quality bottlenecks occur when outputs don’t meet required standards and have to be reworked or scrapped entirely.
If you manufacture a car and the car gets to the end of the line but is deemed unfit and you have to throw the whole car away, that’s like $30,000 in capital that’s now vapor. It’s just gone.
For whatever reason, people don’t translate that concept to code or other business processes, but they should. Consider the simplest feature you could add to your product: you’ve got something that started with a highly compensated person writing requirements. That person spent 30 minutes in a room with several other highly compensated people talking about the requirements. Then two or three highly compensated people turned those requirements into running code over a period of one to two weeks. Other highly compensated people reviewed and tested the feature. And finally, that original highly compensated person decided whether to accept what was built.
So you’ve spent at least $30,000 building a simple checkbox that got added to a form. If that checkbox has to go through that process twice because there was a misspelling on the label, that misspelling just cost you $30,000. (This is an exaggerated example to make the point obvious, hopefully nobody builds single checkboxes this way!)
The quality bottleneck is insidious because it creates waste that often goes unrecognized. When I go into a product development team that can’t deliver code without defects, I sometimes show them a video of an eastern European car manufacturing line from the 70s with a labour intensive quality control process that happens at the end of the manufacturing process. I don’t even comment on it, I just show the video and ask, “So, what do you guys think?” The metaphor is strong.
Friction points
Friction bottlenecks are where the process itself creates resistance to forward progress. This could be user experience friction or internal bureaucracy.
Let’s take a loan application process as an example. One of the big steps is the submission of various documents from the applicant: consent forms for credit checks, tax record access, bank record reviews, and so on. If I don’t make that document collection process super easy, I risk losing the sale because customers get frustrated, walk away, go somewhere else, or just decide they didn’t really need that new motorcycle.
I once had to deal with a mortgage loan process where there was an expanding list on the left with a dropdown box. You had to select a document type, then on the right was a place to upload your document (which had to be a PDF – no other format allowed). You would press upload, wait for it to process, and then go to the next document, doing each one serially. When you were done, there was another button at the bottom that you had to press to actually send all the documents to your file.
That kind of user experience generates a lot of friction. And if your analytics show that you lose a lot of applicants at the document collection step, that user experience is clearly a bottleneck in your revenue generation process.
Internal friction works the same way. If you require three signatures for any change to production, that’s friction. It might have been needed at some point, but if that policy becomes inconsistent with the goals of the organization – like moving faster to meet market demands – then it becomes a bottleneck.
Miscoordination bottlenecks
This is when I think we’re doing X and you think we’re doing Y, and we both spend effort doing those things, but they don’t meet in the middle – like a bridge that’s been constructed from opposite banks of a river simultaneously from two conflicting specs.
In business, this happens when organizations operate in silos without understanding how their actions affect other parts of the system.
Here’s an example: imagine a financial services company that decides to implement another layer of security around their ecosystem in the form of a cloud-based firewall. And imagine that this firewall is provisioned and configured based on nominal historical traffic. Meanwhile, in a separate effort, the company is about to launch a new product with tons of go-to-market activities, including lots of Internet-based marketing communications and responses.
The new firewall crashes, but because there’s no runbook for this scenario yet, everyone scrambles until 2 a.m. trying to restore service. The product launch misses expectations by miles, product managers are angry, IT is blamed… you get the picture.
This is classic siloed thinking, and I’ve seen this pattern play out dozens of times across different industries.
That’s the kind of miscoordination that creates massive bottlenecks and wasted effort.
Aligning bottleneck priorities with your strategic objectives
Here’s the critical question for you as a CEO: where are you trying to go? Do you need to grow top-line revenue by 30%? Do you need to grow your bottom-line margins by 10%? What’s the point, what’s the objective?
Those objectives will determine where you look and the things that become important. If you’re trying to grow top-line revenue by 30%, certain parts of your business might come under more scrutiny for frictions and bottlenecks than others. If you’re trying to reduce the cost of manufacturing by 30%, then bottlenecks in your marketing funnel might not be where to focus.
For example, in financial services, sometimes the mandate isn’t to speed up the loan process but to slow down acceptance of applications so you don’t end up in breach of bank regulations on capital requirements. You want to tighten your criteria when you’re close to those limits and loosen them when you have breathing room. In an optimization problem like this, your most important bottleneck might be the speed at which you can test small changes to your acceptance process.
The point is that your strategic objectives determine which bottlenecks matter most. And that means you need to understand your end-to-end process well enough to identify the constraints that are actually limiting your ability to achieve those objectives.
Taking action: Finding and fixing your constraints
So how do you actually put this into practice? Here’s my straightforward advice:
- Map your key processes end-to-end. For a mid-market company, this might be your order-to-cash process, your product development lifecycle, or your customer acquisition funnel. Whatever directly drives your growth objectives.
- Look for cross-functional handoffs. Companies think in silos. It’s very rare that you truly get cross-functional organized cohesion. But bottlenecks often occur at the boundaries between departments or functions.
- Gather data on throughput and cycle time. Where are things getting stuck? Where are defects being created? Where are customers dropping out?
- Fix the process before you automate it. When Elon Musk built the first Tesla factory for building the Model 3 in Fremont, California, he tried to do it with all robots. They built this entire automated production facility, but when they tried to run it, it didn’t work. The automation process they designed put them way behind schedule, and led to safety incidents and quality issues which took the company to near bankruptcy. They had to shut some of the robots down and hire humans to take up the slack. You’ve got to get the process dialled in before you automate it.
- Focus on one constraint at a time. The Theory of Constraints teaches us that there’s always one primary bottleneck in the system. Fix that one first, then move to the next one that becomes limiting.
In my very first job out of the army, I inherited a very unoptimized production line for a digital wind sensor. I made it my mission to use the Theory of Constraints to streamline that process and automate it as much as I could end-to-end. I reduced the manufacturing time of a single unit from 8 days to three units in one day. That’s the kind of improvement that’s possible when you systematically address process constraints.
Stop masking the pain and solve the root causes
If you don’t understand the process, if you don’t think about the process, and if you’re not looking at the relationships, then your company will continue to think in silos and solve the wrong problems.
The real competitive advantage comes from understanding your business as a system of interconnected processes, identifying the constraints that are actually limiting your growth, and systematically eliminating (or managing) them.
Forget the latest technology or management fad. This is about understanding the fundamental dynamics of your business and addressing the real bottlenecks, not just the ones that happen to be visible or painful in the moment.
As a CEO, your job is to ensure this kind of thinking happens across your organization. Otherwise, you’ll keep taking aspirin for that headache without ever addressing what’s causing it. And eventually, your board is going to find someone who will.