Your data work is part of work politics

Mar 17, 2026

I've been hearing some rumblings of people wanting another DataBS conf? Is that actually true?

Very few people start working and declare that they want to get involved in office politics. After all, workspace politics is often seen as dark and dirty stuff. It's the closed room bartering, the battles about territory, power, resources and influence that make the stuff of TV dramas. And while some jobs are somewhat known up front for being politic-y, like sales roles, others are seen as being more isolated from the direct battles that happen within the upper echelons of management where these things happen – data and engineering being among the latter.

But obviously if I'm bothering to write this post, data work is not particularly isolated from office politics. As the saying goes, everything is political. And while many data practitioners see themselves as relatively neutral actors shining the light of data-driven truth regardless of who's in power, the reality is much different. If anything, my experience is that we're involved, often indirectly, in office politics much earlier than many other office workers because data is seen as having power. And where there's power, there's someone looking to use it.

Data is usually not neutral

A lot of the power of data comes from how most people think that data is a neutral entity that points to truths about objective reality. People with more awareness of the power of data at the very least know that data can be employed to promote one of many viewpoints without doing anything particular unethical or wrong, just by framing questions and analyses in certain ways. But thanks to preconception, everyone who wants to make a persuasive argument wants to have data backing up their point if possible. This naturally translates to stakeholders coming to us as data and analysis providers, to do analyses that further their goals.

Often these are legitimate work requests – if we're trying to build a successful product, then of course we need to define and measure what success means. Even if we know there's a million different metrics to potentially work with, we have to pick something that seems reasonable given what we have, even if we might learn later that our metric was incorrect.

But you don't have to be in this line of work very long to notice that a lot of these "I need a success metrics report" requests comes in waves that are tied to the regular employee performance review cycles. It also doesn't take very long before you start getting requests of "these metrics don't look great, can we find out why? Or slice and dice the numbers in a way to make them look better?" And the reasons for those requests are also pretty obvious.

While my much younger self, fresh out of school and such, didn't notice for a while, these sorts of requests were me being recruited to be a small pawn in the larger game of office politics. When a product manager asks me to do a small side project analysis to help them write a proposal, it effectively means putting my little grain of sand on their scales. Maybe it works out and the PM wins resources to execute their proposal, maybe it doesn't work out. At this stage of the game, I'm still a million miles away from the actual politics, but it was an opportunity to learn what people are looking for, and what they're not looking for. What numbers "look good" (revenue, growth) and what doesn't (just about anything flat or going down).

You can spend quite a bit of time in a career at this level of not having to pay attention to office politics. You can still maintain a relatively neutral stance where you just field data requests in your usual way, stakeholders take the results for their own purposes, and it all just works out in the end. You'll probably get complaints that your results are unfavorable in this or that context, but it's not too much of a burden because at most the requestor will just not use the data you give them.

But eventually you'll have to be somewhat aware of politics

Even if you never try to actively engage with higher level management to request things like resources, and you never try to push for a decision to be made a certain way, at some point in your data career your data work can potentially cause political problems.

The chances of problems happening usually come when you're advancing in your career and your work is being shown further up the management chain. Maybe you start being asked to present to execs. Ever get hit by an executive fire drill where an someone sees a number in a slide deck and asks some questions about the number in a meeting you're not in, and suddenly you lose a week of work because somehow you're the person to go answer that question? Sometimes, it's your number being quoted in a slide deck that triggers that fire drill for someone else. The number was being used (correctly or not) to make a point and someone with enough power to trigger the fire drill called it out for whatever reason.

Every data person who gets sucked into these fire drills usually wants to avoid them as much as possible. You certainly don't want anything you put out to boomerang back into your face. This is where you learn to make sure you don't present numbers that conflict with each other even if they're both "correct" because they come from completely different systems. At the least you put a footnote about it. This is where you learn to shop your presentations around to various people so they can look over and give comment before the big presentation. This is where you learn to be careful with wording and framing and exactly what kind of claims you're making so you're not going beyond what the data can say.

Because even though you're just doing your job and providing the most objective presentation of the data work as you can, the story that you're telling may align, or not, with whatever is going on. It's likely your job to report the objective bad news that revenue is down and the new release was a flop. But for your own benefit, you don't want to appear to put others in a position of having to defend themselves from that data without giving them a chance to prepare a response. It's a fast way to make unnecessary enemies of people who you will still have to work with immediately after.

At some point in all this, you're going to be pulled into political machinations, even if only as an grunt being given orders to follow and being used by others to make a point. I've seen data requests come down from very high levels of management specifically to put certain people on the spot and move things towards a largely pre-determined outcome. I've also been given requests where it's not clear what people want and the data can be interpreted multiple ways so the recommendation from my management chain was to stay as objective and unopinionated as possible so that I don't accidentally throw shade on someone and trigger a whole meeting derailment. Because that has definitely happened before.

Even outside the context of "present data to people with power", you get sucked into politics when you have to set metrics that determine how stuff gets evaluated. How often have you had heated conversations with someone that, when boiled down to the very core, was a variant of "I don't want to use this metric because it makes me look bad"? Even if the metric is 100% the right thing to do, there's a bunch of work to convince the some people to make that metric "The Metric" over the objections of other people. Often that work is the hard part.

Doing well means paying attention

Everything is political, the difference is only in degrees. The fact is that if you're good at doing your data work, then that same work will be used as political tools sooner than later by someone. Maybe it will be yourself, but often it's by someone else. And so when people around you, like your managers, start telling you to take seemingly unnecessary steps like shopping your work around to other stakeholders first to get their feedback, you should be taking that advice to heart. They're usually acting in their own self interest because if you accidentally drop a bombshell into a meeting you're not invited to, they're the ones in the meeting handling the explosion.

Over time, the layers of management between you and the explosions gets thinner and thinner. Learning to minimally navigate these things is a critical "soft" skill. It doesn't mean you have to actively "play the game" in some nefarious "House of Cards" sense. But you at the very least have to realize that your words and analyses have consequences for other people in sometimes un-obvious ways.

It also doesn't mean you have to be a yes-man. Our work isn't just to help our "patron stakeholders" push whatever agendas they want. We should definitely push to show objective bad news when it is necessary because our credibility rests on going where the data leads us. Truth is an important part of our job function. But as my wife often reminds me, how something is said is often as important, if not more important, than what is being said.

Subscribe!

Standing offer: If you created something and would like me to review or share it w/ the data community — just email me by replying to the newsletter emails.

Guest posts: If you’re interested in writing something, a data-related post to either show off work, share an experience, or want help coming up with a topic, please contact me. You don’t need any special credentials or credibility to do so.

"Data People Writing Stuff" webring: Welcomes anyone with a personal site/blog/newsletter/book/etc that is relevant to the data community.

I’m Randy Au, Quantitative UX researcher, former data analyst, and general-purpose data and tech nerd. Counting Stuff is a weekly newsletter about the less-than-sexy aspects of data science, UX research and tech. With some excursions into other fun topics.

All photos/drawings used are taken/created by Randy unless otherwise credited.

randyau.com — homepage, contact info, etc.

All Tuesday posts to Counting Stuff are always free. The newsletter is self hosted. Support from subscribers is what makes everything possible. If you love the content, consider doing any of the following ways to support the newsletter:

Consider a paid subscription – the self-hosted server/email infra is 100% funded via subscriptions, get access to the subscriber's area in the top nav of the site too
Send a one time tip (feel free to change the amount)
Share posts you like with other people!
Join the Approaching Significance Discord — where data folk hang out and can talk a bit about data, and a bit about everything else. Randy moderates the discord. We keep a chill vibe.
Get merch! If shirts and stickers are more your style — There’s a survivorship bias shirt!