Most entrepreneurs are understandably in a rush to get to the right side of the hockey-stick curve (aka scaling).
They do this by trying to go fast on everything.
But going fast on everything doesn’t necessarily make you go faster. It’s a recipe for spreading your already limited resources thin and getting lost faster by focusing on the wrong actions at the wrong time.
Some examples of wrong action, wrong time (premature optimization):
- Trying to optimize a product for thousands of users at the outset -- when you have no or just a handful of users.
- Trying to hire a VP of Sales before you have any customers.
- Trying to raise funding without any traction.
Premature optimization is one of the top killers of startups.
In this issue, I’m going to share practical techniques for focusing on right action, right time.
The key is thinking in systems.
In any system, there is always a single bottleneck or constraint. Think of this as the slowest machine on a factory floor. Trying to optimize all machines is highly wasteful. A better way is to focus on only improving the slowest machine - the limiting constraint.
You might recognize this as the Theory of Constraints (TOC).
TOC is a constraints-driven approach to system optimization, pioneered by Eliyahu Goldratt, and described in his groundbreaking book: The Goal.
Your business model, too, is a system.
Next, let’s cover some concrete steps for how to apply TOC to launching and growing a product.
Step 1. Map your business model throughput (aka Traction) as a series of steps
All businesses share the same universal goal: To make happy customers. We can model these steps as a Customer Factory:
These five macro steps are universal to all businesses, from a flower shop to a complex B2B product. While the macro steps do not change over time, the specific implementation of these steps does change as a product matures.
At this point, don't worry about future rewirings of the factory. Simply map what you're currently doing to implement these five steps.
Step 2. Baseline your Customer Factory
If you aren't already collecting weekly metrics to measure these five steps, start now.
A key attribute of systems is repeatability, and it doesn't take long for patterns to emerge.
If your metrics aren’t consistent across weeks, that becomes your first goal.
Repeatabiilty is a pre-requisite to scalability.
You can't optimize random because random isn't repeatable.
If, on the other hand, your metrics are pretty consistent across weeks, you’ve got a starting baseline and are ready to move on to step 3.
Step 3. Identify your limiting constraint (weakest link)
Working backward (from happy customers), identify the first hotspot in your Customer Factory where you are losing the most users.
It's essential to recognize that identifying this step only tells you what to prioritize (the slowest machine), not how.
Be wary that when you identify a limiting constraint, the common tendency is to want to overcome the constraint by acquiring more resources. For instance, hiring more factory workers or buying more machines.
While these could break the constraint, they can also be needlessly wasteful. What if you could break the constraint by up-skilling your current workers through training or servicing the slowest machine?
To get to the right how you have to first get to the root cause of the poor performance of a particular step.
Step 4. Get to root-cause
In our exuberance for action, we often rush too quickly to formulate and test solutions. But there can be multiple solutions for any given problem, and not all solutions yield the same results. Furthermore, implementing each solution consumes additional resources. How do you find the optimal solution?
The key to finding the best possible solution is taking the requisite time to get to root causes before solution building.
This requires a discovery before validation mindset which opens the door to evidence-based decisions.
Discovery might require further study through customer interviewing, data analysis, or other techniques. This is all time well spent.
The goal of discovery is uncovering insights you can then turn into an experiment and test.
Here’s a quick mnemonic to help:
lean D.I.E.T = Discovery > Insights > Experiment > Traction
Step 5. Make bets on your most promising solutions
It’s certainly possible to still land on multiple possible solutions. Turn each one into small timeboxed experiments and pit them against each other.
You'll quickly know if your solution works from your weekly Customer Factory metrics.
Step 6. Avoid the local optimization trap
Once your metrics improve, don't forget to reassess your overall Customer Factory. As you improve a limiting step, it may no longer be the constraint (slowest machine).
Continuing to improve that step leads to local optimization -- also wasteful.
Constantly search for your limiting constraint, get to root causes, implement a fix, then rinse and repeat.
That is how you turn going fast on everything (aimless wandering) into a more focused and systematic process.
Going slow on the right things helps you go a lot faster in the long run.
