10 Things Learned Making a Web App For 6 Weeks

In July I set out to hack on a web app I wanted to build, called StoryEven. Today is the last day of that exploration.

You can check out the alpha version here @ storyeven.xyz

It’s a super simple web tool that you can use to add headings at the right* spots in your writing. Here it is on Github.

Bookmark it & try it out when you’re editing a long document.

I explain why I built it here.

Here’s a quick brief video on how to use it.

*right = after about every 250 words, at the very least


  1. Building in public works. I would have flaked out due to all the annoying issues I experienced on the way. But because I had announced it, I had to keep going.
  2. 6 weeks is too long. Parkinson’s Law definitely kicks in. I worked on it ~ 1 hour daily. 40 hours stretched into 6 weeks is about 4 weeks too long.
  3. 6 weeks might be OK if you have other things going on. This project also wasn’t exhausting to the point where I had to dedicate all my time to the work on it to finish. Had I set it at 2 weeks, I could have had an (even more) half-baked product.
  4. The platform matters. I picked Netlify for this project because I had wanted to explore it for a long time. Netlify was super easy to use, but my app required Node to run, and Netlify is not meant to run Node (except for build tasks). I compromised and made it work, but it would have been much quicker to set up had I used a different platform.
  5. Git usage goes up on real projects. Up until now, I’ve been hacking on scripts in VS Code. Good for learning JS, but not very realistic for the actual job of building software. On this project, I was able to start using Git daily, and even applying some less well-known methods. Since I was hacking alone, I didn’t get to use all the functionality, but it was definitely better than none.
  6. Vanilla JS is cool, but frameworks exist for a reason. I wanted to do it in as much vanilla JS as possible (still had to use jQuery a bit). But I can see how using React or Next or something similar would have made parts of that work much easier. My project didn’t need their scalability, but their build automation can be helpful.
  7. Don’t start hacking right away. I wrote a PRD which helped me clarify who this was for (NOT bloggers who have many writing tools), among other things. I also made a story map and later refined it further. Those pieces helped me think about the problem and to deliver what I could deliver in the allotted time. A story map was helpful to work through the whole journey of using the product, not only the core of it. It also allowed me to leave out some functionality without giving up on it entirely.
  8. Documentation must be alive. My first story map sucked because I had no clue what issues I would run into. It also didn’t have some elements I later thought of later. My PRD stayed the same, but I am changing it based on feedback from this version. Point is, these documents must live. Your learning must integrate into them.
  9. Get feedback sooner. While I know about the benefits of continuous discovery, I was afraid (and still am!) to share what I worked on. It takes a lot of mental work to be able to do this, and likely why startups don’t do it enough. Knowing about the trap doesn’t mean you avoid the trap.
  10. I root for the engineers even more. While they get paid well, engineering life is not easy. Stuff breaks, stuff doesn’t work, you have to figure out new technology on the go. If you have unrefined stories, you end up procrastinating and any progress is a pain. You also don’t feel like sharing this half-baked work even if though you know you must share it sooner than later.

So, have you tried it yet? 🤑

Idea To Business: 44 Powerful Experiments To Run

One of the key problems experienced by product teams is that they’re not well-versed in the techniques of prototyping & experimenting with ideas.

Not only that, but they don’t often know what options even exist. And there are many! Yet, there’s no one place where one would go to find them all. There’s no body of knowledge that defines what the ways to experiment are.

You don’t necessarily have to A/B test everything, and A/B testing is not often possible. (Besides, experiment design is a serious topic. Many teams hope that it’ll just work itself out when we look at the data.)

However, it is crucial for all businesses and ideas to be tested. It’s the Lean Startup way, and after discovering even a few approaches, I doubt anyone would want to go back to much guesswork.

David J. Bland and Alex Osterwalder came out with a great new book on testing business ideas (affiliate link).

David J. Bland - Testing Business Ideas book cover

It’s my favorite in the series, because it talks at length about all the various ways to discover, test & validate ideas.

This has been my great interest for over 5 years now.

But of course, a book cannot give you a good way to sort through all the various ways, though they try.

I’ve compiled all the experiments, split between Discovery & Validation (i.e. pre-work & post-work).

Below are tests that have the best evidence strength, are the cheapest and have low run times.

Enjoy – and if you want to see more of this data, hit the button below!

Discovery Experiments

Lowest Cost

Top Evidence Strength

Shortest Run Time

Validation Experiments

Lowest Cost

Top Evidence Strength

Shortest Run Time


  • Provide links to practical way to run each experiment
  • Run case studies with each experiment on this site

Recruiters & LinkedIn Don’t Have Problem/Solution Fit? [Product Idea Validation Template]

This post shows you how I validated a product idea for free. I also found out that recruiters have unique challenges.

Below I describe specific steps I took and some techniques and tools I used + how I reasoned about the problem.

At the end, I summarize what I could improve and what you could try instead. 

Why validate a product idea?

You may have heard horror stories about failed product launches or failed companies that folded because of failed products. That’s scary, but you say to yourself: “That’s NEVER going to happen to me.”

How e-commerce platform Elliot fell back down to Earth


All executives become invested in solutions that they promise. They push for their completion without asking “why” or “would anyone care”.

I can commiserate with them. It’s scary to propose something experimental/optional in a culture where everyone expects you to know the one right answer. And yet, you also have to pitch & generate answers all the time (again, culture).

So there you have the perfect storm that creates the famed “Feature Factories”. I’ve worked in them, you’ve worked in them, we all worked in them (and some still do).

There are certainly people who invest into stocks without researching them. People buy courses without researching if other students have eventually reached their goals.

And many products and feature ideas go on roadmaps and backlogs without any validation.

And guess what: a product manager is not immune to this. On the contrary. In fact, I know this from my own professional experience.

Raise your hand if you’ve ever supported a feature that someone on the leadership team had been very passionate about, but had done ZERO market or customer research to confirm if it was something anyone would want?


“The waiter is a product manager who, at heart, is an order taker. They go to their stakeholders, customers, or managers, ask for what they want, and turn those wants into a list of items to be developed. There is no goal. There is no vision. There is no decision making involved.”

(c) Melissa Perri – Escaping The Build Trap

Don’t be a waiter.

a tweet that says "We waste years because we cannot waste hours." by Shane Parrish

The idea

I lost my job a few weeks ago.

Once again, I had CVs and cover letters to write & jobs to apply for. Nobody was hiring, and companies froze lots of roles.

As I communicated with some recruiters, I started to come across people who were less experienced. Some who were abrupt or even rude.

It’s important to remember that everyone is fighting a fight no one else knows about. But it got me thinking how I could improve the experience for candidates like myself.

I wished that there was a way to provide feedback to the recruiters, both good and bad.

Which recruiters were more helpful than others?

I imagined a system that would squeeze over 5000 recruitment professionals in New Zealand to a Top 5 list. That list would be someone’s absolute best chance of getting hired. A crowd-sourced system that required minimal admin.

Not a brand-new idea, but an idea nonetheless.

Still, it felt a bit manufactured to me, and I reasoned that I had to get some feedback.

Before anything else

Write about it

You should write about the first stems of the idea.

Write it in your diary or blog or tweet about it.

I type “docs.new” in my browser (assuming you have Google Docs) and write my thoughts without editing. I write everything that’s piled up in my head: customers, ideas, channels, habit loops – literally anything that my mind’s conjured.

Map it

Next, I use Lean Stack to map my current understanding of this space. I know that it is likely wrong and faulty in many spots at that stage. This is what I need to find out – what are those wrong spots in the first place?

Lean Stack uses a bunch of familiar concepts, one of which is the variation of the Business Model Canvas – the Lean canvas.

Completing it takes less than 30 minutes and it serves as the anchor for the idea from then on.

I recommend Ash’s work to anyone working on new products.

Contemplate risky assumptions

I could see issues with the idea right away. I anticipated a 3-sided product, with recruiters, candidates and hiring companies (clients) all coming to the product to learn about the top recruiters.

There would be problems with motivation, as well as value propositions for each side.

Why would the recruiters want to pay for something that potentially works against them? Only if to improve their score through insights, hidden behind a paywall. That sounded like my main customer would hate my product, right out of the gate.

And if this was a pay-to-participate platform, then recruiters with big pockets would raise their profile, leaving the good independent players in the dust. 

And why would a candidate want to pay for the product? The whole purpose would be to make it public. Would they pay to post reviews of recruiters? That would allow someone with a bit of money to potentially ruin the reputation of anyone on the platform.

My dismay at some unprofessional behavior had generated this idea, so I knew this idea could appeal to someone else.

But imagine disgruntled candidates, rushing to pay for a product that would let them destroy a recruiter who had shunned them.

Then the recruiter would also pay for access to reconcile those reviews? It painted a picture of a very toxic platform, and even the above motivation was hypothetical.

(I also completely missed out on a very obvious solution that already existed. More on that later.)

Still, I wanted to learn about the industry’s problems in general. I set out to explore as I validated my assumptions.

(This post focuses on just one side of that industry – the recruiters – since my first idea was that recruiters would be the primary customers.)


Research plan

I decided to design a low-end research plan. I wrote out my assumptions & questions. Against each, I wrote how and when I will address them.

You don’t have to write a plan, but if you do, IBM has a lot of great resources for this.

screenshot of the exact research plan I used for my first interviews

That anchored me & gave me clarity around ways to verify some of my core assumptions.

Interviews seemed like the most versatile tool. I was eager to do more interviews.

In the past I relied on remote unmoderated studies and surveys, which were safe to do, but limited the amount of insights. Interviews are more intense and can be scary to some at first.

But interviews are worth it, and after this project, I love them and want to do more of them.

Discussion guide

The most important aspect of the interview are the questions. I opted for the topic maps approach. 

screenshot of the clusters of topics made in Miro
several jump-off points

Clusters allow you to stay mindful of topics covered. They allow you to dive into asking about something from different vantage points depending on the conversation flow.


It was easy to source people to interview from LinkedIn. I designed an outreach message and collected a list of people to reach out to. I had Calendly set up and Zoom to host the interviews.

screenshot of a message sent to recruit someone into a study
the outreach
a screenshot of the small Google Sheet where I tracked my interviewees
the spreadsheet

I was the sole interviewer and note-taker, so I had to record my participants.

I don’t recommend you do this alone.

Recording is great even if you’re working with the full team, but transcribing video/audio after the fact was tedious. (I wish I also used Otter.ai to follow along and transcribe in the conversation. Whatever you do here, making sure you’re doing it legally is a good idea.)

Obtain consent for recording. I did this at the beginning of the interview and let the participants know when I started recording. No need for NDAs from my side since there’s nothing to disclose.


Capturing & synthesis

This stage is manual, and I can see why many UX research products have popped up to try to “automate” this.

5-step workflow that I use for analysing interviews, as posted on Twitter

You might have a much better approach and if you feel charitable, please share it!

The first problem didn’t pan out

During my interviews I could see that the “recruiter rank list” idea had no legs. Recruiters didn’t care about their rankings. The industry had no formal governance and the one official organization was a joke to most professionals. They all knew each other through word of mouth. They could see their competitors’ rankings on Google using SEO tools.

I was embarrassed that I missed out on an easy way to give feedback to recruiters. Remember Google Reviews? Yeah, turns out if you google any of the recruitment agencies, you will find the exact information I have been planning to provide. Only it’s free.

In fact, I found my own review for one of the agencies (5 stars to Guy Day at Potentia).

And guess what? There were a lot of bad reviews from people who felt slighted by recruiters’ lack of communication. (This insight is a doorway into our next exploration.)

it hurts…

If not this idea, then what?

I asked a bounty question at the end of each interview.

“If you had a magic wand, what hugely annoying problem would you wish away?”

I don’t recommend this because the insights are not as great as you expect them to be.

People often jump to solutions, fantasise or freeze up, but you can still pick out a few themes. The key is to try to get people to respond with the first thing that comes to mind. You may interrupt if they take too long.

Having those quotes pulled into Airtable or Miro, now you can synthesize.

Don’t overthink it. Grab your notes and arrange them in groups. Name the groups. Do this several times (ideally with others). At the end you’ll have clusters of topics.

Affinity map can be used to validate product ideas by splitting feedback into groups

That’s really it!

What did I learn?

Without revealing much about the participants, I managed to learn a lot about the industry. Being a complete newbie to it, I found out that:

  • Recruiters struggle with time management & communication with candidates. It’s only right: there are many candidates for a role. The recruiter details are often visible on the ad, so people call & email recruiters all day. Trying to balance being nice (someone might be a good candidate) with actually getting things done is a tough feat.
  • The CRMs aren’t always that handy at being the one source of truth. Emails and details slip through, and then you have to do double-entry. It only amplifies that above issue of time management.
  • Competitors are a concern to some, especially when several agencies + internal recruitment team are also involved in the race. Getting a clear read on what contracts to pursue can be a struggle. The solution to this is building trust with clients, usually by sourcing good candidates consistently (i.e. being good at your job).
  • Identifying great candidates is the main challenge. This is what the industry does, essentially. Or does it?
  • People don’t see value in recruiters. They don’t understand what a recruiter does. There is no clear evidence that recruiters indeed assist in hiring people. They may be unavoidable by virtue of being gatekeepers. But most candidates would prefer to talk directly to the hiring manager. This is further supported by the fact that a large part of recruiter’s time is essentially arranging the communication between candidates and the client.

As Hiten Shah, the co-founder of Crazy Egg & KISSmetrics, may put it, recruiters don’t have product/market fit. (Read Hiten’s awesome analysis of why Zoom doesn’t have product/market fit.)

What’s Next?

The 5 themes above are 5 new product problems to explore (that will undoubtedly branch into many more).

For each, we now have to create a new Lean Canvas and explore individually.

This fractal approach is very similar to Melissa Perri’s Product Kata concept. Using it over and over will help validate a number of problems.

An illustration of the Product Kata process by Melissa Perri. It is used to validate product ideas & problems.

What first?

Anchoring the recruiter as the user, I put down their needs and connected them. I could start seeing some connections, similar to what Ryan Singer from Basecamp shared recently, called an “interrelationship diagram”.

A tweet from Ryan Singer with an illustration of the interrelationship diagram method.

Recruiters had an issue with time and communication because they had many candidates reaching out to them.

Those who didn’t cut it, as many don’t, didn’t get a response – which soured the relationship and damaged future trust towards recruiters.

Meanwhile, exceptional candidates rarely (if ever) apply. This, in turn, means that the clients often get second-hand options. This tarnishes their trust and drives the need to hire many recruiters (reasoning that at least one would solve the problem).

the interrelational diagram that helps to decide what problem to start with when validating product ideas
not a pentagram!

New problem

I didn’t want to get into CRM wars. Who has ever heard anyone gushing about their CRM? While this might spell “great opportunity”, going against Salesforce or even Job Adder seemed like a tall feat. It’s also hard to be excited about a CRM.

Time management and email management may have legs. I played with the idea of creating a time management course, or repurposing GTD and selling that. But it seemed like a band aid to apply to a severed hand.

I arrived at the problem of finding & identifying great talent.

But of course, there is the elephant in the room – this is what recruiters do themselves, right? And what about LinkedIn? Isn’t that why LinkedIn exists, to a certain degree?

But ask yourself – do either of these “products” have good fit?

Someone also said that bad candidates would fill out any form you give them, while great candidates wouldn’t even lift a finger. Instead of pulling in all sorts of candidates through an ad, maybe there’s a way to use a push model? A recruiter is really a scout, a talent agent.

My current hypothesis is that there is an untapped market of great hidden talent:

  • Not on LinkedIn at all or is inactive on it;
  • On LinkedIn and active, but not thinking about moving. They’re working in a nice role, have support from leadership and are growing.

These people:

  • Don’t want to fill out any forms;
  • Don’t care if they are unemployed;
  • Are on a career trajectory that is solid;
  • Are so busy that they don’t have time to promote themselves.

How would this affect other problems?

  1. Having great hidden talent on tap would make the recruiter indispensable in the eyes of clients. Some competition doesn’t matter as much.
  2. Time will be spent on priority candidates. It’s like being secure in your relationship instead of playing around. You get a lot of free time.
  3. Communications improve because there are fewer participants (there may still be a lot of back & forth over contract terms).
  4. The funnel is smaller but each candidate is heftier.
  5. CRM data is just a non-issue at this stage.

Next steps

It’s back to the Product Kata. The next step is to validate this new theme.


What I could have done differently?

Interview skills

  • Since I haven’t done many interviews in the past, I feel like having more practice would be beneficial. This was practice though, and I’m glad I did it. But I could have benefited from some coaching.
  • My discussion guide was a mess. I asked some questions and not others. I recalled the clusters and barely looked at it. I only remembered to ask the bounty question at the end – everything else was not scripted.
  • Similar to the first point, when reviewing my questions, I noticed that I often formulated questions poorly. Perhaps I should have stuck with a more structured and rigid discussion guide. Our conversations flowed, but sometimes it went in a direction that added no value.

Lean UX research

  • It would have been good to follow my research plan and do a survey. I would also run some FB/LinkedIn/Google Ads on top of the interviews to gain more quantitative data. Alas, I am in conservation mode until I get my FU money – so I didn’t do those.
  • Subsequently, I didn’t run any smoke tests. I ran out of my Carrd subscription right before COVID-19 hit, and I haven’t renewed. It would have been early to do so to validate anything. However, smoke tests make you think about features, copy and problems in an intense way – and maybe asking those questions would have generated a more robust discussion guide.

If you have any thoughts on how to do things better, hit me up on Twitter, LinkedIn on via email.

If you have used this guide to validate your idea, I would love to hear about it, too!

Run When You See These 7 Red Flags On The First Day Of Your New Job

Red Flags. Hard-to-see patterns that emerge from a mismatch in “what should be” vs. “what is”.

Your new job may look scintillating. The offices are nice, the people are nice, everyone’s nodding in approval.

But it all comes crashing soon.

Maybe you see this on day 1 and you don’t pay attention to it. Maybe these anti-patterns reveal themselves in the next 30–60 days.

Many more red flags have been written about. But I always thought that they were too basic.

Without further ado — the 7 red flags for the new job that you should watch out for…

#1 — They have a giant backlog. It needs your urgent input. It’s a mess.

Not to mix product management and product ownership on purpose, but often companies will use these interchangeably. You can even spot issues at this juncture. Product owners deal with the backlog more. So what role are they hiring for?

In any case, backlogs are often “wishing wells”.

Based on what must one prioritize a backlog?

Why do you have such a giant backlog?

Why are all these features weird, without standard structure to them? Some look like month-long projects. Some would take 30 minutes to do.

Let me guess…

#2 — They don’t do due dates.

Nobody cares if you ship that feature in June instead of April. Or June 2020 instead of June 2018.

They put that feature request 3 years ago. What do you want to do? Park it? Do it?

Ok, why not. Easy, because…

#3 — They don’t know why they do what they do and why anyone needs these features.

What strategic objectives would a carousel slider on the homepage achieve? Why does this part of the flow need a sudden redesign? Who benefits from that thing someone saw on another site and it was cute and now we need it too?

Will the business achieve its goals because of these changes, or in spite of them?

Good questions…

#4 — They hired you to be the bad guy.

Because that’s what PM’s do, right? They say No.

So, start by saying No to this [bad department], because their features look dumb. Did they even do any research?

They want us to build it by April and it is March now and we have tons of work. Who said that the users would buy it?

Can you prove that what they want is bogus, with real research?

#5 — They badmouth other departments.

Yeah, this should be red flag #1.

I mean, you want to work there — but do you need to be in the middle of endless petty feuds? Do you want someone outside the team to try to court you so you could help fight their battles from the inside?

Where are the parents? This feels eerily like high school.

What happens to mean girls? | Stuff.co.nz

#6 — They don’t do buy vs. build. They always build.

Because it’s cheaper. We’ve got all these engineers. For a generalist dabbler like you, writing code is like speaking to aliens. But to these developers, it’s a piece of cake. It’s what they do. D-E-V-E-L-O-P.

Look, no more third-party widgets. It’ll slow the site down. $15 per month? What are you, crazy? We can build it in like zero time. And it’s ours forever.

You know where this is going…

Build the CRM and the A/B testing system and the marketing automation tools and the CMS, etc. etc. etc.

They would probably try to build Google if it didn’t already exist.

#7 — They are so agile, so they scrapped all that agile nonsense.

There’s a difference between (a) testing and keeping the stuff that works and (b) reading about it and then throwing it out completely.

“It wouldn’t work here”.

Yes, collaboration wouldn’t work here, because we’d be arguing about what matters the most forever (See red flag #3). So of course we have to decide by sheer brute force of will and who’s got the biggest 🍆 now!

And who would want to deal with the fact that if we did research, those brilliant ideas wouldn’t stack up? And we would have to rethink, and probably re-do a lot. We’d face the music and the fact that we’re not super heroes. Scary, I know.

And who knew that test & learn was more than just a buzz word and took a lot of work and instrumentation and money and time and alignment?

After all, you might be stuck in that job right now. What can you do about it?

I wish I had a straightforward answer. But that would be incredibly presumptuous.

Frankly, I don’t know a single place that doesn’t have some red flags (I did 7, but that’s just so you would read this piece).

It’s important to remember that no job is perfect, no company is perfect, and no human is perfect. Organisations have complex histories & tend to amplify weird things that don’t represent all individuals in it correctly.

The Best Collection of Great A/B Test Posts

Each section has 1-3 top posts that will help you triangulate and find the common and the uncommon advice on a particular topic.

This video provides a great deep dive into the unpopular truths about A/B tests and statistics involved. Save this for later, it’s 45+ mins.

Let me know if you have unanswered questions in the comments below!

If you ran a test and didn’t get a result

If you have low to 0 traffic…

If you’re looking for a magic number…

If you ran a test and now need to report on it…

The Best Collection of Great Product Management Posts

If you’ve gone remote and worry how your product muscle will work…

If you’ve got a huge backlog and need to make sense of it…

But you’re likely not clear on your strategy…

If you don’t have a product strategy…

If you have an identity crisis as a product person…

If you’re stuck in (tactical) decision-making hell all the time…

Otherwise, pay attention to strategy.

Business Strategy: What It Is & How to Build an Effective One

Strategy Is Hard

If you’ve ever tried to write a clear strategy, read on.

I will dig into the things that might surprise beginner strategists.

Make sure you have ~ 10 minutes to focus. Bookmark it/save it to your tools of choice to read later.

Define Strategy

How do you define strategy?

A strategy is a hypothesis. Based on current intel, certain leverage applied to the problem can fix that problem.

There is always a problem. You generate insights around that problem based on experience and knowledge. Insights help to plan policy and coherent actions. Together they have an exponential effect on the solution of this problem. This is the essence of strategic planning.

The strategy is solving a problem. There are restrictions and tradeoffs. The actions of the strategy solve the problem in the presence of restrictions. They acknowledge the restrictions and work around them.

Here’s one final way to look at it.

Something is already happening in the world. Things that are not in your power; good and bad. You need to know about those forces. You put your best foot forward, in step with what’s happening. A strategy is a direction. Forces bigger than yourself amplify your direction.

Strategy Planning Templates

The first thing you see when you Google “strategic planning” is templates.

These templates guide you down a predictable path of:

  1. Vision
  2. Mission
  3. Goals
  4. Plans

And this simplicity is what makes them so harmful.

Why is it not as simple as the templates make it out to be? Why should you avoid them?

The problem isn’t with the templates themselves. They are correct. They aim to wrap the actions in higher-level abstractions.

The problem is with how you interpret the templates. You think they give you the answer.

You wrote a rosy vision & some fluffy mission statement. You added some unachievable, unrealistic goals. Then you made a list of things to try to do about them. You think, “That’s it, I’m done now.”

But the answer is not that simple to find.

A good strategy is like a symphony, and symphonies are not written in a day.

Why can’t we just call it a day?

Vision and mission are often the outcomes of group-think. They are almost always pure fluff.

Goals emerge from some void if they are not based on metrics you’re already tracking.

Action plans pretend that you know which things correlate with a goal. They pretend that all you have to do is do those things.

But you avoid the uncomfortable question, you avoid the enemy, and so your strategic planning is radically insufficient in formulating a path.

These artifacts of strategic planning lose the point that highlights their worth: they must reflect the strategy as much as the strategy must reflect them.

The vague vision and mission remain vague and drive none of the downstream value of such exercises.

There is a quick sense check you can make.

If your competitors would have roughly the same mission and vision, keep working at it. Don’t your competitors also want to be customer-centric? Don’t they want to grow 20x the market rate of growth? Yes, they do.

So what’s your strategy?

Strategy Is Design

You have to know what’s going on right now.

What is the current state of things? What is and what isn’t, to the best of your knowledge?

Research is context-specific and is difficult to do. And someone has to do it.

Start by asking questions.

What is your problem, and what is the challenge with it?

The unspoken secret is that good strategy is almost always a surprise, but it simultaneously rests on very sound basic business principles. And simple is hard.

Giving people something they need is basic advice.

How to give it to them? How to know what they need? How to do all that in the face of competition, in a way that works for the business model and the economics of it?

You need research. You need an outside view. You need to synthesize everything together.

You need to write out the symphony as it plays out.

To a designer, these elements start to look familiar.

A good strategist is a designer at heart.

The work is never done…

You may have identified the problems as you see them, and they are right, but you have no leverage to work on them. So are they your problems? Or is the real problem — the lack of leverage?

You are not looking for the most polished strategy in the world. I am not advocating an endless session of navel-gazing.

If you don’t have an insight, do work to get it, because your problem is “no insight”.  There’s nothing embarrassing about it. It’s more embarrassing to rely on a strategy document that is a waste of paper.

If You’re Stuck With Your Strategy

Try these approaches:

  • Restate the problem
  • Gain some outside leverage
  • Pause and research
  • Act, to find out what is happening.

You do not need a brainstorming session. You should even avoid one, especially in a group setting. You must find out what’s happening, not seek false hope in someone’s well-sold idea.

Your strategy is a hypothesis, and to test the hypothesis, you need prototypes, and testing and experimentation. Don’t sit in a stuffy room, pretending to do work. Don’t create 100-page presentations. Nobody reads them! Test it.

What you should have now is:

  • A problem with a unique set of challenges;
  • An insight and, hopefully, some leverage.

So how do you translate this into action?

Strategy Actions: Coherence & Conflict

One reason why we often choose conflicting actions is that we want to hedge our bets.

Of course, you must keep BAU happening, but you must keep it happening without making BAU the main strategy, or you soon won’t have much of BAU to worry about.

What companies usually do is that they try to optimize every single aspect of the company by X%. This is how businesses interpret the concept of the balanced scorecard.

In analytics, there is a statistical procedure called the Principal Component Analysis.

It’s a useful technique that decomposes a dataset into a smaller dataset that has all the main elements tightened down to as few as possible. If you had 50 columns, now you have 3.

These components are made up in such a way to explain the variance of the whole dataset in the most efficient way. You don’t have to worry about 50 variables, just 3 principal components now.

But these 3 components are arranged in such a way to avoid loss of information.

So if there are variables that tug in one direction and other variables that tug in the other, they will be expressed separately as 2 components. They have to explain the whole field of choices.

But in truth, you don’t need the whole field of choices, which all offset each other as they cover the most ground.

You need just 1.

The balanced scorecard is not the 3 components.

It should be all the original variables that go into the 1 component that your strategy will be based on.

You’re betting on 1 number, not hedging by betting on all 3.

Or at least, the part of your strategy that bets on the other 2 only kicks in if the original strategy didn’t work.

Strategists speak of focus, but the manager thinks that focus means several things at once, done rapidly and with brute force.

So what are the components of the policy and actions?

  • The policy describes the main principle of the strategy;
  • Actions are concrete things that specific people will do according to the policy;
  • Actions must be checked for coherence (do they supplement each other or pull resources apart?);
  • Actions must be checked for conflict (do they cancel each other out?).

What About Goals & Objectives?

They’re not enough for a strategy, and they’re not a replacement for strategy.

Once the actions and policies are clear, goals and objectives will easily emerge. You can now put them on a dashboard because you have a list of actions that will be triggered if goal A is not reached, or if KPI X goes down.

Goals may stay the same, but objectives can change. Increasing sales by 50% can be done by a huge variety of ways.

And goals may change based on the policy.

Don’t put the carriage before the horse (or the donkey 👇)

📑 Summary

I cannot promise you that this post has solved your critical problems around building strategy.

Here’s what you should take away today:

  • Templates for strategy formation breed fluff and bad strategy. Don’t rely on them alone — think!
  • Group-think and brainstorming may hurt your strategy. You need an outside view.
  • Goals are not substitutes for strategy, they don’t come before it, they are not dictators. Objectives may change based on goals. Don’t start with them, but keep them in mind.
  • By avoiding the enemy — the problem & the restrictions — you forfeit the whole exercise. Don’t be lazy, don’t be dismissive of it.
  • Leverage is the core component of the strategy. This is the one thing that, if missing, invalidates the strategy completely.
  • Actions the policy generates must not be incoherent (tugging in various directions) and conflicting (cancelling each other out);
  • None of this is guaranteed to be successful!

Your strategy is still a hypothesis.

Questions About Strategy

How do you understand the limitations of a strategy and when to get back to it?

If a core team member is sick, how do we make sure we achieve the goal?

The strategy will always have limitations. You have to make choices. It’s deceitful to think that the best strategy can achieve it all. How to understand the limitations? If you have made choices, look at the cutting room floor, find the pieces you have decided not to do: those are your limitations.

If a person is not available, can someone else do it? Can you hire quickly enough?

This is more tactical and your strategy shouldn’t disintegrate because of such questions unless this person WAS your leverage and the whole strategy was to hire THIS genius to do the work.

If they WERE, you should have protected that leverage with everything in your power.

If they WEREN’T, then they were not as crucial to the fulfilment of the strategy and you can find a way to go without them for a while.

How do people present strategy?

Most companies post their strategies online. Of course, there are internal documents which may be hard to reach.

Then there are the pitch decks to investors, which are a good place to start.

Some companies explicitly state their strategy on their blogs. You can find examples below.

To create your own, a pitch deck seems like the perfect starting point, but a well-written blog post also works. Think memos.

  • If it’s too long, nobody will read it.
  • If it’s too short, it’s fluff.
  • If there’s a chart, make sure it has a purpose.

A strategy cannot be so simple that it fits in a small sentence, but it shouldn’t be a book either.

What if the best strategy to solve a problem may not fit the culture of an organisation?

If the culture fit wasn’t there, you probably went in with a template strategy and didn’t think well enough, i.e. you had no strategy for this particular client.

Entering a new environment should always start with the research and the understanding of the problem.

How can you fix it now? What is this thing in the culture of the organization that is preventing the “best” strategy from working? What are those limitations? Is it certain people who oppose it? Is it the guidelines from the top? Are there time limitations, or technology limitations, or skill limitations on your side?

A strategy is a hypothesis. Your roadmap is the strategy, and it’s not guaranteed to win.

If you understand the problem well enough, now that you’re in the org, do your best to diagnose the problem. There are some specific tools that can help unearth more insight into the problem. That’s not the point.

Examples of bad strategy


Whenever you hear stuff like “continue to focus…”, “further strengthening…”, you know there is fluff coming up.

Take a look at this chart from the corporate strategy document posted on the website of Yokogawa (Japanese electrical engineering and software company) website:

Key strategies for a new mid-term business plan:TF2020

Does that tell you anything? What’s digital transformation, and why do it?

Perhaps not all of it is fluff.

Perhaps some parts of the plan are there to alleviate certain investor concerns.

Perhaps the real strategy is elsewhere and it’s really good, but why post this?

[Source: https://www.yokogawa.com/about/company-overview/corporate-strategy/]


In one statement, the company commits to probably conflicting goals of simplification + addition of a new business unit.

Cost reduction may or may not provide an additional layer of complexity to the simplification policy, because to simplify, you probably need to do some digging, and then probably some automation, unless you’re cutting costs everywhere — which doesn’t recognize the competition, the market, the landscape, and what will happen once you do improve profitability.

And here’s a deceptively organic flywheel-looking chart that actually fails to communicate how one thing leads to another:

Our strategy

[Source: https://www.telstra.com.au/aboutus/our-company/future/Ourcorporatestrategy]


The pinnacle of fluff is Hilti’s statements on strategy and goals.

“We passionately create enthusiastic customers and build a better future.” Fluff.

The focus areas say nothing about the challenges, just BAU, or it’s cousin, “Try-Harder BAU”.

[Source: https://www.hilti.group/content/hilti/CP/XX/en/company/corporate-information/Strategy/corporate-strategy.html]

Examples of good strategy

The Tesla Master Plan

Superhuman + How Superhuman Built an Engine To Find Product Market Fit

What is Coinbase’s Strategy?

The first pitch deck by AirBnB founders

Uber’s Strategy by Andrew Chen