3.6: Data Toddlers

The CG only had a couple hours before his engagement that morning. The topic of discussion was Army force structure cuts, as impactful an issue as any I worked on in the headquarters. He had a few clarifying questions which were quickly answered and then he asked two new, unanticipated queries that would require another look at the data itself.

If I’d had my laptop with me, I could have answered them in the room, but like an idiot I’d left it in my office. This was in part because staying connected to data wasn't how our headquarters worked. Minus a couple token ports in a few conference rooms, there were few places to connect in the headquarters to the internet or even power.¹ We weren’t setup to be connected to data, because we didn't work in data. Honestly, it was generally looked down on to have your laptop in meetings. That was for back seaters, not for directors and primaries. Our culture was one where if anyone worked in data, it was your staff or your assistant, not you.

I stepped out into the front office and asked the CG’s team to pull up the spreadsheet that had been given weeks earlier so I could run the new numbers. They struggled to find the file. No one had taken the time over the previous week to look at it. Interacting with data wasn’t what these fellow lieutenant colonels did either. Not that it would have mattered. They were better at PowerPoint than Microsoft excel and weren’t going to be able to adjust the pivot table to rerun the new math. They were data toddlers who'd been given the keys to a sports car; they couldn't have done anything with it if they wanted.

I sprinted back to my office and made the adjustments, which was contrary to expectation as well. I was a director, so I was supposed to kick it back down to the deputies who’d gathered the data in the first place. That was how our headquarters ran.

But that was too slow. The CG needed answers within the next 45 minutes, and I was already kicking myself for the 15 I’d lost by not having my laptop in the meeting.

Our culture, one where it’s not the commander or director’s job to know data, is what prompted me to write my first article, the one that spawned this entire substack. I firmly believe this culture of keeping data at arms distance from a commander is the main impediment to broad data literacy across our army.

This culture starts from a good idea, but it devolves along the way. At the army’s pre-command courses, they stress ‘Do only what you, the commander, can do’. This isn't entirely wrong. You have a team of subordinate commanders and a staff, and you should absolutely leverage them, empower them to go after data and to make their own decisions.

In the case of the force structure data, I’d done just that. Chastity, my deputy, had compiled the massive dataset for me after I'd given her the details I wanted. You readers can probably guess that compiling a dataset for me can be a real pain in the ass. I often ask for a lot of ancillary data so I can setup lookups across the data sets. It's a lot of work, but thankfully just the kind of challenge my deputy enjoyed.

Just because a commander delegates compiling and managing data to a subordinate, doesn't excuse them from understanding how to use it. This is because one of those things ‘only the commander can do’ is ask commander questions. I’ve previously written about how a commander needs data literacy to form good questions (Footnote: See 2.09, 3.3, 3.4). But there’s no excuse for a commander to expect only their staff to interrogate the data to find answers.

It takes too much wasted time to wait on your entire chain of command and staff officers to slowly roll a task back down a hallway of other data toddlers until it gets to the one staffer who can rerun numbers for a new question. By rerunning the numbers myself, setting up a new pivot table fairly quickly, I could respond faster to the CGs RFIs. Chastity, who was always busy, could keep doing her real job as a Force Manager. I could do the work myself because I’m comfortable with working in data.

Being data literate isn’t just about moving faster. It’s also about reducing risks. It means you can tell when the data is wrong. If you as a leader aren’t data literate, you leave yourself vulnerable to those who do understand data. You must take their word for it. It's all too easy for the data arrogant to take advantage of you. I know this because I've done it.

Back in my young lieutenant days when I was a company XO in the infantry, our CSM tasked the companies to compile an incredibly detailed tracker for where every soldier would be on their Christmas leave, well before everyone had finalized their plans. He demanded the document before Thanksgiving, and defiantly claimed no leave would get approved that wasn't turned in by his suspense. I did try to remind him that CSMs don't approve leave, commanders do, which did nothing to make him like me more.

There's no point in arguing with a CSM who lacks the authority he thinks he has, so instead I opted to get him off our company's back. I cheated. I compiled a list of every soldier in the company, and input the data we did have, then just made up the rest with my best guess. Next, I added a bunch of spurious data to the table. Details that could be relevant but were just there to clutter the view: location of leave, phone number, platoon and squad. Then I added more variables that were completely irrelevant: blood type, social security number, PT score, etc. Next, I added vibrant color bands to each by the type of leave and finally I sorted the whole thing alphabetically. The resulting spreadsheet was painful to look at, which was the whole point. The CSM took one glance at the wall of data on the PowerPoint slide and blanched before remarking, 'Well it looks like Charlie Company has this well in hand' and moved on to harass another company.²

I watched a similar exchange happen two decades later in the lead up to that morning with the CG. The whole army was facing cuts and ARSOF was right there in the thick of it.³ My force management counterpart at SOCOM was strictly analog. He was particularly prone to complaining about how the math on his PowerPoint slides wasn't adding up but refused every call to stop using PowerPoint for what absolutely needed to be a data-driven decision.

The army was certainly using data. They'd provided a detailed report from the Center for Army Analysis that argued for very steep cuts to ARSOF. When the SOCOM commander had demanded to see the data that justified them, the Army was all too happy to send the raw data. They knew SOCOM couldn't read it, never mind do anything with it. And SOCOM didn't.

Frustrated, I asked to see the army’s data. It wasn't well annotated, nor was it in table form.⁴ But after poking at the data for about an hour I knew enough to call up my Pentagon counterpart. 'Your numbers total to X, but I thought they should sum to Y.' Then, offering up a guess I asked, 'Is this why?'

'Not many people catch that,' he replied. He then took the time to help walk me through the outcomes of the analysis they’d done and help me understand both where the army was correct, and where ARSOF hadn’t given them the right starting data. I could attack the numbers we'd been given now.

I could do it because I didn't have to ask a subordinate to ask a subordinate to ask a subordinate to redo the numbers. I could move at the speed of the data. But I could also argue with another person’s data because I could poke it and find the holes in it. Commanders need data literacy as much for their own self-reliance as to ensure they aren't going into a meeting with nothing but PowerPoint-deep talking points. Especially when the other side is coming in with data. When they do, your side is going to lose because it’s constantly running back outside the room for answers. You cannot respond on the fly.⁵

Circling back to those ‘data toddlers’, lieutenant colonels working front office jobs. One thing toddlers do is watch and learn. They emulate those around them, especially those who look like they’re in charge. It’s in our DNA.

The army puts some of its brightest officers in these front office / bag man jobs: XOs, ADCs, CAGs or CIGs. These are the officers assessed with the highest potential for future command.⁶ They’re working these jobs instead of being directors who are charged with moving the needle on force design or innovation. The army does this because we want them to get exposed to the world of general officers. But they’re also seeing the example set by their principal. Too often it’s a commander who is largely data — and digitally — illiterate. Commanders who model the learned helplessness every day to the next generation of field grade leaders.

And the cycle begins anew.

1. You want more drops in your conference rooms. Trust me.

2. For once I was happy this was on PowerPoint and there was no way to interact with the data.

3. ARSOF - Army Special Operations, which includes SF as well as CA (Civil Affairs), PO (Psychological Operations, or PSYOP), the Rangers, and ARSOAC (Army Special Operations Aviation Command. Special Operations Command, the head of all US special operations in the DoD.

4. ‘Format as table’ should become one of your best friends in Excel. It gives you slicers, and makes formulas markedly easier not just to write, but for others to read.

5. Two months earlier, I sat outside the ASD/SOLIC office, laptop in hand, to help prep them for the engagement with the army. After waiting for thirty minutes, I was told I wasn’t getting into the room, but I could try and dial in from a VTC suite somewhere else in the building. They spent an hour going over SOCOM’s PowerPoint slides, trying to come up with answers that weren’t on them. They were in the data on a laptop outside the room.

6. Commander’s Action Group. Also called a (CIG) Commander’s Initiative Group. Ostensibly a mini staff for general officers that works prep and projects for the CG. IMO? Stolen manpower from the staff.