AnalyticsHacker.com

We often say things like:

• Let’s optimise conversion rate
→ we’ll get more sales

• Let’s increase email CTR
→ it will prove our CRM strategy works

• Let’s expand to new markets
→ it will make us look good externally

On the surface, these sound sensible.
They’re measurable.
Actionable.
Impressive.

But this is often the same kind of thinking as saying:

• Let me drastically cut calories
→ I’ll lose weight

• Let me double my exercise volume
→ I’ll see faster gains

• Let me read X books per month
→ I’ll become smarter

Will you see results quickly?
Probably yes.

But will it be:

• Sustainable? → probably not
• Aligned with what you actually need? → often no
• Healthy, ethical, or grounded long-term? → unlikely

Because optimisation is easy when you don’t question the goal.

As AI makes optimisation dangerously easy, the real skill (and responsibility) is knowing when to STOP, reframe, or say no.

And this is where things usually go wrong.

Not in ex*****on.
Not in intelligence.
But in what we choose to optimise for.

We tend to prioritise:

• Short-term wins
• Metrics that signal competence
• Ego-driven incentives (“this will show progress”)

And we quietly ignore:

• Second- and third-order effects
• Lagging indicators (trust, loyalty, resilience)
• Team fatigue when work feels performative, not meaningful
• Whether this effort actually moves us closer to the right outcome

So what should you do instead?

Slow down — before you speed up.

• Interrogate the problem framing, not just the solution
• Ask what success looks like after the metric moves
• Separate “this is measurable” from “this actually matters”
• Create space for uncomfortable questions — especially the ones that threaten the plan
• Stop trying to rationalise every decision
→ you often know more than you can tell or prove (a lot of your knowledge is experience-based, difficult to articulate or codify)
→ trust that knowledge
• And yes — have some fun with it

Good decision-making doesn’t start with optimisation. It starts with sense-making.

With acting —> noticing —> adjusting.
With allowing understanding to emerge while you’re engaged — not waiting for certainty that never comes.

And that’s the scary part:

You can be very good at solving the wrong problem — and still fail.

Speed doesn’t replace clarity.
More data doesn’t fix a flawed frame.

Better outcomes usually don’t come from doing more…

They come from choosing the right problem to solve in the first place.

You’ve probably been a thinking partner this week without even realizing it.

That moment when a colleague talked through a problem and suddenly said, “Wait, I think I know what to do”?
You didn’t solve it for them.
You created space for them to think.

This is thinking partnership—and it’s one of the most underestimated tools in decision-making, in life and work.

Here’s what the science tells us:
Our brains think differently when we’re truly heard. When someone listens without judgment, our brain gets the space it needs to work. But when we feel pressured or judged, we shift into defensive mode, and our best thinking shuts down.
Articulating thoughts out loud helps us organize complex information. It’s called “self-explanation”—and it only works when there’s a listener creating the psychological safety to explore ideas without fear of being wrong.

What makes someone a great thinking partner?
It’s not about having all the answers. It’s about:
• Active listening without jumping to solutions (because your silence creates space for their insight)
• Asking incisive questions that help them examine assumptions they didn’t know they were making
• Creating a judgment-free environment where independent thinking can emerge
• Trusting that the other person has the insights they need—they just need space to find them

For decision-makers:
The next time you’re stuck, don’t just ask “What should I do?”
Ask someone to think WITH you.

The clarity often comes not from their advice, but from hearing yourself articulate your thoughts to someone truly listening. You’re literally using their attention to think more clearly.

For thinking partners:
Your job isn’t to fix, advise, or direct.
It’s to listen deeply and resist the urge to rescue.
The question “What do YOU think?” is often more powerful than any answer you could personally give.
Why? Because you’re signaling: “Your thinking matters. I trust your capability.”

The everyday magic:
We do this unconsciously in many of our conversations—with friends making career moves, teammates wrestling with data interpretations, or leaders navigating complex decisions.

The magic happens when we do it intentionally.

People who talk through decisions with a good listener (not an advice-giver) report higher confidence and satisfaction with their choices. They own their decisions because they truly made them.

Try this week:
In your next one-on-one or coffee chat, practice being fully present.
Listen more.
Advise less.
Ask the question that helps them see what they already know.

Notice what happens when you trust someone else’s thinking process.
You might be surprised how often the “breakthrough” was always there—just waiting for the right conditions to emerge.

Routine analytics should be automated.

Humans should not spend their time querying dashboards.
Pattern detection is not insight.
Speed and scale belong to machines.

Most analytical labour should disappear.

But not all insight originates in data.

Some of the most consequential inputs to decision-making sit outside what can be measured or optimised — including ethical judgment, an understanding of human behaviour beyond observed metrics, emotional and cultural context, second- and third-order consequences, and responsibility for harm, not just performance.

AI can optimise within a frame.

Humans must choose the frame.
Challenge whether it is valid.
Decide when not to optimise.
And carry moral and organisational responsibility for outcomes.

This is not a defence of human analysis.
It is a defence of human judgment.

The more we automate ex*****on,
the more judgment matters.

For a long time, the analyst’s role was clear.

A question arrived.
Analysis followed.
An answer was produced.

That rhythm shaped how organisations learned to think.

Now answers arrive faster than they can be absorbed.

They become cheap.
Abundant.
Often immediate.

What remains difficult is knowing which questions deserve attention.

That difficulty now sits upstream of analysis.

Vision and strategy are often described as data-driven.

Which is usually true — right up to a point.

Data is very good at showing what is happening.
It reveals patterns in what already exists.
It sharpens our view of risk, constraint, and trade-off.

But it does not tell you why any of it matters.

Vision does not emerge from analysis.
It is a commitment that exists before measurement has anything to work on.

When vision is confined to the data already available, it quietly contracts.
It stays close to what has happened before.
To what can be measured now. To what we “believe” is possible.
To what the system already knows how to see.

Many of the decisions that actually shape direction don’t live there.
They sit ahead of the data.
Often the data only appears once the decision has already been taken.

This is where many businesses begin to stall.

They ask for more evidence to support a direction that cannot yet be justified.
They refine what is known while postponing the choice that would change the frame entirely.

Data still matters.
It can inform.
It can test.
It can unsettle.

But it does not originate strategy.

And when strategy starts to feel reactive or constrained,
it’s rarely because the data is insufficient.

It’s usually because vision has been asked to stay within the limits
of what is already measured.

At a certain point, more data stops improving the decision.
Not as a failure of analytics — but as the edge of analytics.

It’s often the moment when:
∙ teams start circling the problem
∙ something important feels unspoken
∙ the discomfort of choosing without certainty becomes visible

More dashboards are built.
Another analysis is run.
An alignment meeting is scheduled.

And yet — the decision doesn’t resolve.

Because sometimes the hardest part isn’t the data.
It’s naming the real question.

And sometimes it’s recognising that not everything that matters can be measured.
That doesn’t make it less real.

This is usually where the work shifts — from information to judgment.
Where clarity comes not from more analysis, but from reclaiming human responsibility in the decision.

That is the space I work in.

Most teams are rewarded for answers, not questions.

The cost is rarely visible at first.

Don’t be fooled by the “average.”
Especially if you’re a Marketer or Business Leader.

We love a good summary stat:
• “The average customer spends $120”
• “Our average email open rate is 24%”
• “The average user session is 3 minutes”

It sounds clean. Digestible. Easy to report.
But here’s the catch: the average can be wildly misleading.

Let’s say:
One VIP customer spends $1,000, while nine others spend $10.
The average? $109.
How many people actually spend that much?
Zero.

When we build strategies around the average, we risk targeting no one in particular—just a mathematical illusion.

⸻

What to use instead:
• 📊 Median – the middle value, less affected by outliers
• 📈 Distribution – how are values spread? Are there peaks or gaps?
• 🧠 Segmentation – group customers by behavior, spend, lifecycle

⸻

The average can be a starting point,
but real insight lives in the details, the deviations, the outliers.

So next time someone throws out an “average,” ask:
• Is it truly representative?
• What story is the data really telling?
• Are we oversimplifying something that needs more nuance?

Because smart strategy doesn’t chase the average—
it learns from the edges.

🔎 Data analysts, looking to make your work a bit smoother?

AI tools are here to help with tasks like automation, visualization, and finding insights faster. This blog lists some of the most useful tools for analysts, including Tableau, ChatGPT, Power BI, and more. 📊🤖

If you’re curious about practical ways to enhance your workflow, take a look!

🔗 Read the blog here: https://www.analyticshacker.com/analytics-resources/best-ai-tools-for-data-analysts

🚀 Just released my latest YouTube video: "AI Coding Assistants for Data Analysis & Workflow Automation"! 🎥

In this video, we explore how AI coding assistants like GitHub Copilot, Codeium, and Gemini can transform your coding workflow, whether you’re working with Python, R, SQL, or other languages for data analysis. These tools not only help you write code faster but also improve your coding skills, enhance productivity, and assist with real-time debugging and documentation. With those AI tools, you will be able to write code even if you are a complete beginner.



AI Coding Assistants for Data Analysis & Workflow Automation

AI Coding Assistants for Data Analysis & Workflow Automation Introduction to AI Coding Assistants for Data Analysis | GitHub Copilot, Codeium & Gemini DemoIn this video, we explore how AI coding assistants like GitHub ...

Working on a new course for you guys. Follow me on YouTube for latest content.

💡 25 Ways to Use ChatGPT as a Data Analyst 📊