The Data Daily
Less than 2 minutes to read each morning.
Not sure if you want in? Read the archives below.
5 days a week since May 1st, 2023.
How and why to hit the brakes.
Brakes get a bad reputation.
Bikes, scooters, cars, and ATVs all have brakes.
Brakes help you slow down or stop.
They feel protective. Restrictive. Limiting.
They hold you back.
And when that's the story, then why would you ever want brakes for
- Your data team?
- Your career?
- Your business?
Here's why you want brakes.
Brakes are about agility and managing complexity. The majority of the time you use brakes is not to stop, it’s to adjust to a complex environment around you.
When riding a mountain bike on a dirt single-track trail, I rarely use brakes to come to a stop. But I constantly tap brakes to gain control and navigate through tight, sharp, steep, rough, or uncertain sections of the trail.
Overall, my brakes allow me to go faster, with better control, and a higher success rate.
But if you think of brakes as something that will slow you down, stop you, or get in your way you will be looking for ways to remove or avoid brakes.
Rather than embracing them.
What do brakes look like on your data team?
- Pull request approvals, deployment gating, or code reviews.
- Pipeline failure process and auditing.
- New data request prioritization and justification requirements
- Continuous Deployment of code changes to new environments
The tasks, processes, and habits that help you interact successfully with complex environments - either code, people, or data - are your brakes.
They aren’t there to slow you down. Or stop you.
They are there to help you go fast in a challenging world.
What do brakes look like in your day-to-day on your data team?
I’m here
Sawyer
Cohorts and Bootcamps - Learning and growing with data this summer
In this episode of Making Data Matter, Troy and I share about two different data learning experiences we've been a part of. Troy in a technical bootcamp and Sawyer with a data leadership cohort.
We discuss:
Tech stacks, open-source tools, and big data
Learning experiences and methods
Leadership vs. individual contributor problems and challenges
and more.
That’s a bad idea
Have bad ideas.
Write bad code.
Make bad art.
Why would you ever want to do that?
Every now and then you will stumble on something good.
Slowly those “good” things collect.
And occasionally something great comes out.
If your data team isn’t allowed to fail, they rarely will succeed.
If your data team doesn’t have the safety to produce a host of bad ideas, they will rarely have good ones.
The goal isn’t to create bad stuff.
It’s to not be scared of makeing something bad.
I’m here,
Sawyer
A cure for metric madness.
A pandemic among dashboards and analytical reports is something I call metrics madness.
Building measures, charts, graphs, and cards for every. possible. metric. that the developer or stakeholder can think of.
The report expands with data points like a balloon.
Until it eventually pops.
When your executive looks at a dashboard they need less noise. Less clutter. Less confusion
Try this minimalist framework.
Only two types of metrics are allowed: Terminal Metrics and Decision Metrics.
Terminal Metrics are end-goal metrics. They aren’t progress markers, rather they represent the outcome of the organization. Revenue. Patient Outcomes. Number of children kept safe.
Decision Metrics are metrics that you track because they are tied to specific decisions. When the conversion rate drops to Y% we change the marketing campaign. When program attendance increases by X% we add staff or a new site.
Here’s the important point - you ignore every other metric. Every other number. Every other data point.
If it’s not a terminal metric or a decision metric it’s not on a dashboard or reviewed by leadership.
The end result?
Much less clutter.
Far more clarity.
I’m glad you are here,
Sawyer
They don’t want your bottleneck.
You can’t remove a bottleneck.
You can only move it.
With your data team you will encounter bottlenecks constantly.
Some are technology.
Some are budgetary.
Some are process.
Some are human.
Bottlenecks are a constrained part of a system the restricts the movement and production.
It’s the narrow point.
If a VP has to approve every job offer before it can be sent to a candidate, the VP is a bottleneck in the hiring process.
If you only have one employee skilled at building Power BI reports, they are a bottleneck for report requests.
At times you will need to address a bottleneck because it’s too restrictive for your organization to function well.
Here’s the crucial idea to embrace. Bottlenecks don’t go away. They only move.
Always know where you are moving it.
Many times we move the bottleneck out of our sight and then pretend it’s gone. Pipeline running too slow? Scale up server to a higher tier. Did you delete the bottleneck? No, you might have just moved it the balance sheet or the QA process.
If someone is pushing back on your efforts to remove a bottleneck? Pay attention to why. It’s likely you are moving the bottleneck to something that impacts them directly.
The greatest data leaders constantly think about the systems, strategies, and constraints of your team and org.
What’s a bottleneck you are facing right now?
I’m here,
Sawyer
How to have good fights on your data team
…Making mistakes.
…Trying hard things.
…Empowering creativity
…Aiming for ambitious goals.
…Interacting with other teams.
…Attempting things that may not work.
All of these things are healthy for a data team aiming to make real difference in their organization.
And they are also a common environment for conflict to arise.
Yet, not all conflict is created equal. Knowing what’s healthy conflict and what’s damaging to the team is challenge I’ve often wondered about.
This comment on LinkedIn gave me a lot of clarity.
“…There are two kinds of conflict - (1) affective conflict where you don’t like the other person and (2) cognitive conflict where you don’t agree with the ideas of the other person.
Cognitive conflict is good and necessary and provides for a better solution to the issues that the team/organization is facing because a number of different alternatives are placed on the table and debated and discussed.
Whereas affective conflict is bad because you don’t like the other person and that prevents you from sharing information or withholding information because someone you don’t like is leading the discussion or proposing a solution, that could help the organization.”
As a leader, your job is to encourage cognitive conflict. It’s stimulating and healthy for your team.
At the same time, craft ways to reduce affective conflict. It’s toxic for your team.
If your team is going to fight, make sure they are fighting about the things in the right way.
The most important data challenges aren’t about technology or data itself.
It’s about people.
I’m here,
Sawyer
How to hire data talent the wrong way
Right now, we filter candidates first based on technical skills. You give them a coding challenge, a take-home project, or quiz them about their tech stack experience.
Only later, likely in the final phases of the interview dance, will you explore human and business skills. Asking questions to understand how they solve problems, can they express empathy, and if they have an innate curiosity about others. And do they “fit” with the team.
This will flip in the next 5 years.
The technical skills will be commoditized as tooling and AI advance. Their previous experience or knowledge of SQL will be decreasingly relevant.
Rather, you will be far more interested in their human skills.
Do they know how to articulate a challenging situation?
Do they show an appreciation for the challenges of the business teams?
And, perhaps most importantly, do they know how to identify and embrace objectives?
AI will increasingly make it easy to do things.
So many things.
What it won’t be good at (perhaps ever), is telling you want.
The objectives of a company, organization, team, or individual won’t be set by AI. That’s a human challenge.
Or to put it a bit clearer.
AI will be great at the how and what.
It can’t give you the why.
When you are hiring, you will have to search for the “why” people, because those who excel at the what and how will be a commodity.
I’m glad yo are here,
Sawyer
A ball and chain around their ankle
Nobody has time for that.
Your leadership doesn’t have the time
To wade through a crowded dashboard.
To decide in the moment which data points are relevant for their decision.
To parse through multiple pages of an Excel workbooks to understand a formula reference.
When you deliver a report to your leadership filled with noise you might as well give them a ball and chain around ankle.
The most valuable assets a data team can offer are:
Clarity
Confidence
Decision-Centricity
Empathy with the challenges your leaders face is a fundamental part of being a data professional.
And it’s the first step to creating actionable, clear, and simple data solutions.
I’m glad you are here,
Sawyer
Your data is expensive if you aren’t doing this.
Should we collect that data?
Should we build that dashboard?
Should we integrate with that other data source?
Here’s a simple two step process for deciding:
Define what decision that data will serve.
Define how that decision serves your end outcome.
“I want to know how we are doing on sales” is not a decision
“We are curious about our turnaround time for support tickets” is not a decision.
A decision is an action.
“When new leads from our marketing campaign falls below XYZ a month we will drop the campaign” is an action.
“When a supplier's on-time performance is above XY% we shift more business to their account” is action.
Collecting data is a liability.
It’s costly, painful, and time-consuming to develop.
And it’s costly, painful, and time-consuming to maintain.
Make sure you know why you are collecting it. And what you will do with it.
I’m here,
Sawyer
Your skill at making decisions
“Data-driven” is not primarily about data.
Data-driven does not require
TBs of data
Large data team budgets
Sophisticated data infrastructure
Expensive data literacy training initiatives.
Data-driven actions can happen with very small amount of data, with immature data technologies.
Because being data driven is not primarily about data. Data-driven is an approach to making decisions.
It’s about clearly identifying decisions.
Quantifying the uncertainty or risk in a decision.
And using data to increase the likelihood of success in an economical way.
Sometimes that requires a lot of information (data). Sometime it requires very little.
Sometimes that demands massive cloud-scale systems. Sometimes all it needs is Excel.
If you are going to measure anything in your company, you should be measuring your skill at making decisions.
And data is the most effective tool to help you make better decisions.
I’m here,
Sawyer
That’s how daring ventures start.
“The more effectively dependent people are on one another the more independent and daring they become.” - Amir Levine
You might think of creativity as a risky act. And it certainly is. A fundamental part of creativity is making something while knowing ‘this may not work’.
You might think independence is risky. And it certainly is. Taking a step off by yourself to something new, different, or unknown is unsettling for the nervous system.
But the best parts of life exist in a beautiful tension.
—> You are most risky when you feel secure in your community, team, and partnership.
—> You are most independent when your attachment to your tribe is strong and unthreatened.
—> The safer you are in your relationships the riskier you can be.
—> The more secure you are in your community the more creative you can be.
What does this have to do with data?
Humans are best at work when they are secure, risky, and creative. The type of work you most desperately need your data team to do is risky and creative.
Helping a teammate get unstuck.
Following the data wherever it leads.
Admitting they made a mistake on a project.
Choosing organizational goals over their own.
Pushing back on long-established assumptions.
Stepping into a difficult conflict with peer leadership.
Trying a new narrative to explain the complexity of the work.
Being courageous to prioritize the work that matters most to the team.
These are the actions you want on your team.
The prerequisite you can’t skip? Strong, safe, and secure relationships.
That’s how daring ventures start.
I’m here,
Sawyer
What is a data-driven decision?
What is a data-driven decision?
-> A decision that has less risk or uncertainty because of data.
The fundamental nature of running an organization is making decisions. That's what executive leaders do day in and day out.
Their decisions are what they are hired for, fired for, and get bonuses for.
What strategy do we adopt?
What location should we open next?
Who should we hire for this important role?
How much should we invest in this new project?
All of these decisions are filled with risk and uncertainty. You don’t have perfect information about the world.
The more certainty you have about a decision, the more likely you are to get the outcome you want.
The primary value of data is in reducing the uncertainty in these decisions.
People want outcomes.
They are constantly making decisions to try and get those outcomes.
If you can make their outcomes more likely (using data), then they will sing your praises.
Promotions, career growth, budget approvals, and beyond come easier after that.
I’m here,
Sawyer
The only metrics that you optimize for.
Yesterday, we talked about the challenge of not knowing where you are going.
Today, is the second (and more common challenge) data leaders face.
You don’t know how to get there.
Some data leaders I talk with have clear visions of what the future could look like for their team. You’ve done the work.
But when you look ahead
Or when you look behind at the progress you made
You see:
Frustrating delays from conflicting organizational goals
Technology debt that requires more maintenance than it delivers in value.
Disappointing response from peer teams that aren’t interested in your goals.
Apathetic engagement from your team members on new initiatives.
Budget approvals that go on for longer than you’ve put off cleaning out your garage.
Another team member resigns and leaves you shorthanded.
The gap between where you are and where you want to be seems to be growing.
Your confidence about what to do next is shrinking.
You don’t know how to get there.
But you aren’t alone. Start here:
Look at your future goal. This is your end destination. Then identify and define two metrics that would track your progress. Most of your struggle with progress is due to noise and confusion of swirling complexity.
One destination. Two progress metrics.
This dramatically simplifies your work. These are the only metrics on your team that you optimize for. Now, you can focus your efforts to only activities that will move those metrics.
The rest is noise.
If your road is still cloudy, and you can’t figure out what metrics make sense, I have a Measuring Success Launchpad offer to do exactly this. It helps you define peak success in detail to get you from the present to the future state. I can only take on 2 clients per month. Hit reply now to get started.
I’m here,
Sawyer
From support tickets to peak success
There are two core challenges you feel as a data leader. Here’s the first (the second will be tomorrow)
When someone asks “how is work?” how do you respond?
“Busy”
Most people look at their last week, and the week ahead and just think “busy”.
You and your team are doing lots of stuff. New dashboards are happening. An occasional new data source is integrated. Some code refactoring. Performance tuning. Bug fixing here, there, and everywhere.
See? Lots of stuff.
“Busy”.
But are you confident it’s the right stuff? Is “busy” really what you want?
Are you moving the needle on what matters? The stakeholder teams ask for things all the time, and you try to accommodate them as much as you can. Yet, it still feels like missing alignment and constantly shifting priorities are your daily life.
Here’s the challenge - you don’t know where you are going.
You are going everywhere at once and nowhere at all.
How do you move from a firefighting data team to a data team the organization can’t live without?
How do you get out of support tickets and bug fixes to purposeful and strategic data initiatives?
You get clear on where you are going and why.
With technology, it’s easier than ever to do lots of stuff.
But harder than ever to know what do to.
Or why you should do it.
If you want to do this on your own, here’s your plan:
Clear an afternoon from your calendar. Get away from your desk, step out of the office, or at least away from your inbox and phone.
A notepad and pen will be all you need.
Start with 3 simple questions:
What’s the main marker of success for us?
How will we know when we’ve gotten there?
What decisions do I face when making progress toward success?
Take two hours and sit with those questions only. If you give your mind the real space it needs, you will have no problem filling up a page or two of notes.
If you can’t make this happen on your own, I partner with data teams to get clear on how to measure and reach peak success. It’s a two-week sprint. Designed specifically to give you confidence and clarity.
I can only take on 2 clients per month. Hit reply now to get started.
I’m here,
Sawyer
Can’t stop the feeling
Everyone’s ultimate desired outcome is a feeling.
They say they want more revenue. But they want to feel like they are secure.
They say they want higher customer satisfaction. But they want to feel like they are doing meaningful work.
They say they want a promotion. But they want to feel like their work is appreciated.
They say they want improved child safety outcomes. But they want to feel like they are making a difference for others.
They say they want increased users or members. But they want to feel like their mission matters.
Safety. Confidence. Recognition. Appreciation. Success. Contentment. Fulfillment. Great leaders identify these types of drives and desires from their people and their leadership.
A key part of your work as a data leader is
1) Knowing what feeling you want.
2) Knowing what feeling those around you want.
If you can make people feel how they want to feel you will never lack for success, fulfillment, or opportunity.
I’m here,
Sawyer
The other side of the coin from courage
Fear goes with courage like peanut butter with jelly.
“I’m scared” is just on the other side of the coin from courage.
Courage is a virtue. Fear isn't. But they always dance together.
Two ways to face this week. One for you and one for someone else.
Make something that scares you. Write, draw, speak, share, dream, listen, step, stop, hug, or hold. It doesn’t matter how big the action or the creation is to others. It matters that you saw your fear but turned the coin over and saw your courage.
Help someone be courageous again. When you see someone’s actions and you think “That took guts”, call out the fear that they had to dance with to take that step. “I bet that was scary. I see the courage that took”. Hearing someone else acknowledge your courage is huge validation and gets them a step closer to acting courageous again.
Your work in data is moving more and more away from commoditized actions (writing code, arranging reports, etc.) and into more deeply human acts. Courage, curiosity, compassion, creativity and connection.
I’m glad you are here,
Sawyer
We get stuck yelling about cloud security, data models, and excel
Getting someone to “buy-in” has very little to do with the thing itself.
Rather it has to do with purpose and desired outcomes.
See if this sounds familiar:
—> Scene 1: Director of Data – “Hey we are moving to a new analytics and reporting tool. Check it out, it will be great for you”. Stakeholder – “No, I’m good. I like the old one and Excel best”.
—> Scene 2: Data Architect – “We are moving to a cloud data platform. Over the next few months we will be retiring all on-prem databases”. On-prem DBA – “That’s the stupidest idea I’ve ever heard. The cloud is way too expensive. You are inviting major security liabilities.
—> Scene 3: Data Modeler – “We are moving to dimensional modeling practices across all of our business units”. Finance Manager – “Nah, please just continue to give me everything in a flat file”
—> Scene 4 (from the Nyquist house): Parent – “Alright, it’s time to get ready for bed.” Kid – “No”.
In all of these scenarios, you could make an argument for why one side is an overall better decision. But “overall better decision” is entirely based on what outcome you want.
In scene 1 the stakeholder’s desired outcome might have nothing to do with which reporting tool they use. Their core outcome is to have confidence in their ability to do their job. A new tool threatens that.
In scene 2 the DBA has a core outcome of keeping their job and feeding their family. The cloud isn’t what matters. Their job matters.
In scene 3 the Finance manager's primary focus is on being able to deliver end-of-month reporting accurately and effectively. Changing the format of the data could take her hours longer next month and she just doesn’t have that kind of time.
In scene 4 bed times means no more playing. Playing is fun. Fun is the only goal.
In all these scenarios, you will fight a losing battle if you argue over “the thing”. If we aren’t sharing an objective then we will get stuck yelling about cloud security, data models, or what time the clock says (bedtime). None of those things really matter to the other person.
Shared outcomes are the only way to get buy-in.
I’m here,
Sawyer
I’m washing my hands of this
Pre-s: Last call! Today is the last day to register for the summer cohort of Technical and Strategic Data Leader.
If you’ve waiting for the right time to join, it’s here.
——————-
A while ago I worked for a while on a report for a client. It had a bunch of formulas, some complex SQL and some unique ways they needed the data presented.
When I sent the email to the team that requested the report, I was proud. I was confident it did what they wanted and they would be thrilled.
They replied a few minutes later thrilled to begin using the new report. It would save them hours of time during their end-of-month reporting.
Ahh. I thought to myself. Time to wash my hands of this one and move on to my next challenge.
Until they emailed me a week later with a bug they found.
Then again the next week with a revision request.
Then again three weeks later with a discrepancy in the numbers.
I was annoyed. They took my beautiful report and all my hard work and they started poking holes in it, changing the requirements, and gasp finding bugs.
Why was I annoyed?
I had embraced a broken theory of partnership. In my mindset, I had “won” by delivering the report, regardless of how useable or useful it was to the stakeholders. Mentally I had moved on from this report and considered a win which I delivered my part.
Underneath my annoyance was an attitude that believed I could be successful without the users being successful.
I could win regardless of whether they win.
But the partnership is the flip side of that. I win when you win. And I don’t win if you don’t win.
That fundamentally changed how I viewed that report and the stakeholders who kept emailing me. Now, I was in it for their success. And I’m not successful until they are.
I’m glad you are here,
Sawyer
Data Leaders can’t ignore this trend
LLMs are an essential part of the future of data.
Yes, there is a hype cycle, but the underlying technology and the pace of innovation are still game-changing.
Data Leaders can’t ignore this trend, nor can they be ignorant of what it means for them and their team.
What are the core things a data leader needs to know about Gen AI?
How does Gen AI fit into or interact with my data architecture?
What are practical use cases for Gen AI in our organization?
So we are adding a bonus 7th Week to our Data Leadership Cohort to talk about just this topic.
My friend (and industry expert) Hitesh Govind will be leading our conversation on the topics, showing real demos of Gen AI with customer use cases, and answer questions.
If you were on the fence about joining our cohort, now’s the time to jump in.
This is your last chance to take a huge step in your data leadership journey this summer.
Tomorrow is the last day to register.
I’m here,
Sawyer
Practical tips for designing success
No guest interview this week on the podcast - It’s a solo episode with me. This is an audio recording of a recent presentation from a Linkedin Webinar.
You can find the video version of this presentation here.
This episode covers:
Why is measuring success for data teams so hard?
What are the consequences of doing this poorly?
What are the results when data teams do define and measure their success?
A practical framework for defining success metrics.
Tactical tips for designing metrics
Examples of how this works in practice.