The importance of importance - “There is always a most important thing. If you don't know what the most important thing is you will not be important for very long.”
/By Duncan Anderson. To see all blogs click here.
One sentence summary:To maximise your success you need to understand what is the most important thing and construct your first principle view on what should be done for it
If you are running a team it is likely not possible to know everything that is going on, even if the team is 3 people… let alone 30 or 300.
IMO the manager of a team must always know what the most important thing is, and have a personally derived first party first principles view on what to do for the most important thing.
IMO it is not ok for the manager of a team to not know what the most important thing is.
IMO it is not ok for the manager of a team not to have their own first party first principles view of what to do for the most important thing.
The manager likely cannot know everything, but IMO the manager must know these things.
Also, even if you are not a manager, I think you should know what the most important thing for yourself is and how to improve it if you have been in your role for 6 months or longer.
Jingle: “There is always a most important thing. If you don't know what the most important thing is, you will not be important for very long.” TM DA
“If you want to become important then get good at figuring out what is important!” TM DA
Importance doesn’t lead to impotence, it leads to importance, ah hahah!
Personal Story:
Modus operandi - 5 years ago: I used to just try as hard as possible.
Doing a good job = trying as hard as possible.
Let’s call this working hard but not necessarily smart.
Modus operandi - now:
Work to figure out what the most important thing is and then what I think should be done for the most important thing.
Please note: the most important thing isn’t necessarily the biggest thing.
Let’s call this ‘trying to work smart!’
Modus operandi - future:
I don’t know yet but i change my mind about basically everything so assuming I’ll change my mind about this.
Why do you need to understand what is important
You should always be trying to improve at work
Understanding where to improve requires understanding what is the most important place for your time
As a manager you need to understand what is the most important area for your team
But more than that you need to understand what is most important for you, your projects and anything you are doing at work
By understanding the importance you will understand priorities
This will lead to you being much more effective and successful in the work you complete
How to know if you understand what is important?
I like the analogy that what you are doing at work is ‘building a machine’. It might be just the work you are personally doing, or it might be for a team, or an entire company, or an economy! Basically you can approximate anything as a machine.
Here are some problems I have seen with corresponding questions to help you understand whether they are problems for you :
People don’t know what the machine that they are operating looks like => can I make a flowchart of the machine schematic at sufficient detail?
People have an idea of what the machine looks like but no idea of how well each node in the machine is functioning => do I know what the key metric that matters for each node in the machine is?
People don’t have a view on what is the target level of performance for each node => do I have a target for each metric based on first principles of what is possible? (no ‘the metric should improve year on year’, sufficiency vs perfectionism, etc)
People don’t know what the most important thing for themselves / their team is => if you have the above pieces you should be able to determine the most important part of the machine.
People don’t have their own first party first principles view on what a good solution is for the most important node (no outsourcing this one) => you don’t have to have a view on how to do everything well in every part of your organisation but for the most important thing I believe you should.
+++++++++++++++
Delectable details:
Theory:
Why do you need to know what is important?
At work you should always be working on improving. Be you a brand new recruit straight out of university / school or the CEO of a large corporation.
If you are going to improve, you need to understand what should be your top priority to improve
“If everything is a priority, nothing is a priority.” Typically once your workout what is important focus on this until it is at sufficiency and then move onto the next thing to improve!
Almost everything can be improved indefinitely, so at some point you stop and reallocate your time because it is not the most important thing to work on anymore.
Another lens to look that this through is where can you get the most ROI with your time
Improving sufficiently… or improving sufficiency… or sufficiently improving? How about all three!
How to figure out what is important?
I put on my special counsel hat and do an ‘important investigation’.
I like the analogy that what you are doing at work is ‘building a machine’. It might be just the work you are personally doing, or there might be for a team, or an entire company, or an economy! Basically you can approximate anything as a machine IMO.
The process below might seem a bit confusing at first but hopefully it will make of sense once you read the examples
The “important investigation” process… AKA the machine that improves machines, ha:
1. Map your your problem space end game.
I.e What does the optimal solution look like
2. Map out the theoretical maximum path to your problem space end game.
I.e what is the best case scenario for getting to your optimal solution
This could be adoption rates, time to get there etc
3. If you are already achieving close to your theoretical maximum don’t dive deeper, everything is all good and therefore this is not an important issue to understand! If you are not close to theoretical maximum go a layer deeper and try understand what is important.
4. Map out your existing machine in a flow chart. (you are effectively building a model of your problem space)
“All models are wrong but some are useful.”
Don’t make mangey models, make magnificent models!
5. At each node (A node can be any discrete part of the flowchart for your machine that can be analyzed) of your machine, determine if it is:
5.1 A place in the machine where a theoretical maximum exists or not.
5.2 if it doesn’t have a theoretical maximum determine the minimum sufficiency needed for the node.
6. Determine existing performance at each node vs theoretical max or sufficiency.
7. Determine which node is the most important to work on in the existing machine design taking into account the consequential nature of nodes
ie there might be a node down the stream that can improve more in absolute terms but it’s actually more important to work on the upstream node because of flow on effects
8. Now that you have a high level understanding the existing machine you need to decide what to do next. My typical options:
8.1 do I need to go deeper on a specific node, as currently I only have a low resolution understanding that is insufficient to make a decision about prioritising what to do
8.2 is it better to look at changing the machine design (ie the way the flow chart flows ;) )
8.3 and if neither of 8.1 or 8.2 make sense then get into trying to improve the node deemed most important!
How do you improve a node once it is identified as being important?
Say I’ve identified a node that I’d like to try to understand and / or improve, how do I normally look to break down the node? This is one model I sequentially run through (but normally I custom make up a model / taxonomy that tries to fit the problem I’m solving)
The model:
Recipe - the detail of how to do whatever problem at high quality
Process - how to do the problem at high quantity (if your recipe is sh1t, don’t try and scale it (process). No one wants lots of sh1t.)
Can you do it straight up faster
Can you do a small, medium or large version of what we are currently doing
Should we remove a piece of the process
Should we add a piece to the process
Is there an out of the box solution to making things that is a big win.
Resourcing - is there enough time/money to get a sufficient outcome?
Culture - are the people doing the work aligned with each other and the business?
Team structure - Does the team have the right people, the right roles and the right people in the right roles?
Comment:
There is no point working on process if the recipe is poor IMO.
It doesn’t matter if you have great resourcing if they are making a bad reference design (recipe) in an inefficient and / or error prone way (process).
Because you have a great recipe, process and underlying human resourcing doesn’t mean you won’t have a counter productive culture… but you can’t have good culture if people are crap, they don’t like the recipe or the process does everyone’s head in.
Don’t get me wrong, any of the pieces above being wrong will wreck something, but I normally find it best to go sequentially through pieces in the manner I’ve just said.
Edifying Examples:
Example - Simple Enterprise Sales Organisation:
Background:
I’m going to over simplify this example to try and make the model work easily.
Product: You make timetabling software for secondary schools (ie what student should be in which class with which teacher)
Currently schools do this by pen and paper and your solution is both faster and cheaper
You also don’t have any competitors in the software space.
1. Map your your problem space end game.
What is the theoretical maximum market share of your product? 100% as it’s faster and cheaper than the existing pen and paper solution.
2. Map out the theoretical maximum path to your problem space end game.
What is the theoretical maximum adoption curve of your product?
Map out all the valid reasons a school could say no (eg need to see other schools using the product, eg can only deal with so much change at a given time, etc etc).
This might mean that in the first year you can get a maximum of 20% of the market, 40% in the second year, 60% in the third year, 80% fourth year and 100% in the fifth year.
3. If you are operating close to theoretical maximum don’t dive deeper, everything is all good! If you are not close to theoretical maximum go a layer deeper.
How is the machine performing currently?
You are about to enter the 3rd year of selling your product
In the first year you took 10% market share, and in the second year you took another 10% market share for 20% market share in total.
So vs your theoretical maximum you are underperforming by 50% (ie should be at 40% market share by your estimates but are only at 20% market share).
So we go a layer deeper.
4. Map out your existing machine in a flow chart. Ie you are effectively building a model.
So you build a ‘flow chart map’ for the machine to see where you can improve things
Existing machine map:
1. Figure out who the right person at a school to speak to is
=> 2. Phone call to set up meeting with person
=> 3. In person meeting
=> 4. School makes / doesn’t make purchase of the product
5. At each node of your machine, determine if it is:
Ok, annoyingly I’m going to jumble the next steps together and just synthesise.
Sample machine map 1:
Synthesis:
IMO the clear step to look at here is that only 20% of meetings are going to purchase.
There isn’t much point in closing more than 50% of phone calls unless we are confident that 20% is the maximum you can get out of meeting people.
So you make a map of the meeting step to see what you can do?
1. Product is much worse than we think
2. We are pitching the product with poor sales materials (ie recipe)
3. The sales people have good materials but aren’t using them well
4. The school isn’t ready to make a decision at this point, ie after 1x meeting.
You do the work and find out that it’s ‘reason 4’ and you then change the sales process to include a trial.
Proposed sales machine after this round of work:
You then run the machine to gather enough data to re-examine the machine and see if things are near theoretical maximums.
Sample machine map 2:
Synthesis:
Annoyingly you have only moved from 9% overall close rate to 10% despite the new steps.
However you’ve gone from ‘in person meeting’ of 20% close to 50% onto the next step. So this looks like it is a good win.
To me the glaring problem to me here is that after a trial there is ‘only’ a 50% close rate after a school has tried the product.
There are two major possible problems I see here :
1. The product isn’t anywhere near as good as you believe it to be
Or 2. The trial is run poorly and the appropriate experience isn’t had.
You investigate for the possible areas here and find out that while the product is faster and cheaper it is only if schools know how to use it. You need to make a bunch of User Experience improvements for the product, it’s not how the training is being run, it’s that they are being trained to use a product that is unnecessarily hard to use.
You make improvements then run the machine to gather enough data to re-examine the machine and see if things are near theoretical maximums.
Sample machine map 3:
Yay, you have moved to 18%, an 80% jump.
This is close to what you had thought was the maximum amount of customers you could close, 20%. Should you stop here as this from a low resolution point of view is where you felt conversion rates might cap out?
You have learned a lot about your sales machine, and you look at the ‘in person meeting’ conversion rate of meeting => trail of 50%. If the product is really good this should be higher right? What are the reasons why this number is so low:
1. The product is no good => you shouldn’t have 90% close rate after the trail if the product is no good.
2. The sales message in the meeting isn’t doing a good job of explaining the actual product
3. The people in the meeting have a quality sales message to convey but are not doing it well.
You investigate and find out that the sales message isn’t at all capturing and resonating with the appropriate school personnel. You do the work to upgrade the sales message and then run the machine to gather enough data to re-examine the machine and see if things are near theoretical maximums.
Sample machine map 4:
You get the in person meeting => trial of software number to go from 50 => 80%. Fark yeah!
Overall close rate is now at 29%.
This is more than 3x where you started at with your sales machine!
This also means your thoughts on theoretical maximum adoption rate was wrong.
You might decide that these numbers can’t be improved upon and then go about scaling your sales team, eg you had a team of 5, there wasn’t point in scaling a team that wasn’t working well, but now it’s time to go to 20 and you then have the problem of how to scale the team as the key thing to focus on!
Example - Running a restaurant:
Background:
I’m going to over simplify this example to try and make the model work easily.
One of the key things I’m going to try and flex in this example vs the previous sales example is there were ‘theoretical maximums’ in many places, here there won’t be.
You have a restaurant running and you have the meals that you make. Let’s see what is going on.
Why is a restaurant a good example to you?
To me, in some respects making product is like designing a meal (ie making recipe)
Do you have a good meal (recipe) that people will like?
You can do a minimum viable product (MVP) and test out a new recipe on people before you scale it to your entire menu?
Once you have a recipe right can you get the right ingredients for it?
Can only one person make the dish or can you get many people to make the dish?
1. Map your your problem space end game.
The goal is to have a thriving restaurant with 80%+ occupancy each week.
2. Map out the theoretical maximum path to your problem space end game.
Unlike selling timetabling software to schools, a restaurant can actually start with high occupancy and stay there.
Eg you start out at 80%+ occupancy and stay there.
Eg you start out at 80% occupancy and then it starts to drop
Eg you start out below 80% occupancy and occupancy rises from there.
For this restaurant occupancy started at 50% and has been steadily dropping to now being at 30% occupancy.
3. If you are operating close to theoretical maximum don’t dive deeper, everything is all good! If you are not close to theoretical maximum go a layer deeper.
This is obviously not close to theoretical maximum. We need to figure out what to do.
4. Map out your existing machine in a flow chart. Ie you are effectively building a model.
Flowchart:
1. The recipe
2. The ingredients for the recipe
3. The cooking of the recipe (process)
5. At each node of your machine determine if it is:
5.1 A place where a theoretical maximum exists or not.
5.2 if it doesn’t have a theoretical maximum determine the minimum sufficiency needed for the node.
Analysis
1. The recipe => there is no such thing as ‘theoretical max’ for a recipe. ie you can always do something better. This is like saying a book is as good as a book can be.
2. The ingredients for the recipe => while i’m sure there can always be better ingredients, this an area where a theortical max is more possible than a recipe having theoretical max IMO. It’s like half way between ‘has a theoretical max and doesn’t have one’, ha!
3. The cooking of the recipe => this one I feel there is a theoretical max.
6. Determine existing performance at each node vs theoretical max or sufficiency.
You are the owner of the restaurant, and you determine that when you cook the meals you have very happy customers.
However not so for your other chefs.
So the problem is “3. The cooking of the recipe” (process, scale)
7. Determine which node is the most important to work on in the existing machine design taking into account the consequential nature of nodes (ie there might be a node down the stream that can improve more in absolute terms but it’s actually more important to work on the upstream node because of flow on effects)
As above, “3. The cooking of the recipe” is the area to focus on.
8. Now that you have a high level understanding the existing machine you need to decide what to do next. My typical options:
8.1 do i need to go deeper on a specific node as currently I only have a low resolution understanding that is insufficient to make a decision about prioritising what to do
8.2 is it better to look at changing the machine design (ie the way the flow chart flows ;) )
8.3 and if neither of 8.1 or 8.2 make sense then get into trying to improve the node deemed most important!
Possible solutions for “3. The cooking of the recipe”:
Is it because you have chefs who aren’t trained properly?
Is it because you have bad chefs who aren’t trying properly?
Is it because the recipe is so complicated that only a master chef can do it?
Synthesis
You know any decent chef could make this dish as you are personally cooking.
The way to find out if ‘you have bad chefs who aren’t trying properly’ vs ‘you have chefs who aren’t trained properly’ is to try and train the chefs and see how this goes. This is how you figure out what the root cause problem is
You conduct this project to get your data and then synthesize!
Ok, enough examples! What have I seen?
People don’t know what the machine that they are operating looks like => can I make a flowchart of the machine schematic at sufficient detail?
People have an idea of what the machine looks like but no idea of how well each node in the machine is functioning => do I know what the key metric that matters for each node in the machine is?
People don’t have a view on what is the target level of performance for each node => do I have a target for each metric based on first principles of what is possible? (no ‘the metric should improve year on year’, sufficiency vs perfectionism, etc)
People don’t know what the most important thing for themselves / their team is => if you have the above pieces you should be able to determine the most important part of the machine.
People don’t have their own first party first principles view on what a good solution is for the most important node (no outsourcing this one) => you don’t have to have a view on how to do everything well in every part of your organisation but for the most important thing I believe you should.