FinPod

[00:00:00:02 - 00:13:11:17]
Imagine just for a second that you are sitting in a high stakes Monday morning leadership meeting. Oh yeah, we've all been in those. Right. The coffee is entirely too strong. The tension in the room is just palpable. And everyone around that massive mahogany table is staring up at a giant screen. Waiting for the numbers. Exactly. Have you ever noticed how the numbers you focus on in those specific moments end up completely ruling your life and your decisions? It's kind of terrifying when you really think about it. Whatever is up on that screen, whatever metrics are flashing in front of you in green or red, they are basically about to dictate your entire week. Because as we'll see, those screens aren't just innocent displays of data. No, not at all. So today we are pulling from a couple of really fascinating sources. We've got excerpts from the executive dashboard, lessons in corporate governance and metrics, and also the corporate finance institute's guide on KPIs in FP&A, which is financial planning and analysis. Yeah, the FP&A teams are the ones usually building these things. Right. And our mission for this deep dive is to figure out how finance teams and leaders take these absolute mountains of raw data and somehow forge them into actual coherent decisions. To do that, I mean, we really have to start by completely reframing how we think about this tool. Okay, how so? Well, an executive dashboard is not some boring static report. It is fundamentally an attention allocation engine. A attention allocation engine. I like that. Yeah, think of it as a highly sophisticated piece of behavioral design. It shapes human behavior because the numbers you choose to put on that screen literally dictate what your team will optimize for. Right, because that's what the boss is looking at. Exactly. And conversely, the numbers you leave off that screen, they just fade into the background. They essentially cease to exist in the minds of your team. Out of sight, out of mind. For better or worse, yeah, the dashboard is the absolute loudest voice in the room. So before we get into how massive, complex organizations actually deploy these things in the real world, we need to establish what a true dashboard even is. Because people mix the terms up all the time. They really do. In typical corporate jargon, people tend to use the words report and dashboard interchangeably, but functionally, they are entirely different instruments. They serve completely opposite purposes. Think about it this way. A report is comprehensive. It exists to give you all the granular details, much like a patient's entire medical chart. You sit down, you pour a cup of coffee, and you read a comprehensive FP&A report for an hour to really understand the deep history. But a dashboard isn't that. No, a dashboard is curated specifically to focus your attention in the immediate present. It is not the medical chart. It's the pulse oximeter clipped to your finger. Oh, that's a great way to look at it. Yeah. It's just giving you the vital signs to tell you if the patient is crashing right now. Precisely. A highly effective executive dashboard generally contains roughly seven to 15 metrics. That's it. Only seven to 15. Yeah, just a handful. And its entire architectural purpose is to allow a leadership team to answer three specific questions in about 90 seconds. Okay, what are the three questions? First, how are we doing right now? Second, where are we surprised? And third, what requires a decision from us this week? Wow, in 90 seconds. Yeah. If you're staring at a screen and you cannot answer those three questions in a minute and a half, you are not looking at a dashboard. You're looking at a poorly formatted report that's basically wasting everyone's time. Okay, let's unpack this. Because if you only have seven to 15 slots available on this highly coveted piece of digital real estate, how do teams avoid filling them with just useless numbers? It's tough. It happens a lot. Right. We have all sat in meetings staring at a dashboard filled with metrics that look impressive. Maybe a chart going up and to the right, but they don't actually tell you anything meaningful. It requires an extreme level of operational discipline. Our sources from the Corporate Finance Institute actually outline four non-negotiable characteristics of an effective KPI or key performance indicator. Okay, lay them out for us. So they have to be relevant to the actual stated strategy of the business. Makes sense. They have to be measurable with reliable hard data rather than like subjective guesses or surveys. They have to be simple enough that they don't require a five-minute footnote from the finance team to explain. Nobody wants to read the footnotes in a meeting. Exactly. But the absolute most critical characteristic, the one where most companies fail, is actionability. Actionability. Let me jump in with an analogy here. Because understanding the danger of non-actionable metrics, which are often called vanity metrics, is so crucial. Go for it. Tracking an Altoon without an actionable input is exactly like standing on a bathroom scale that only tells you that you're heavy. Right. It gives you a number, sure, but it doesn't tell you what you ate yesterday. It doesn't tell you how many calories you burn. It doesn't track your cardiovascular exercise. It's just the final result. Exactly. It just gives you this final output that you cannot immediately alter. If a metric goes in the wrong direction on your corporate dashboard, and your leadership team's only response is, "Well, we just track it. We can't really control it," then that metric has absolute no business being on your dashboard. What's fascinating here is how often companies fall into that exact trap. All the time. They track the school of the game without tracking the specific plays that generate the score. I mean, if you cannot tie a specific immediate action to a metric, you are basically reducing to a spectator of your own business. That is a great way to phrase it. A spectator. Those numbers belong in a quarterly retrospective, not on a weekly action-oriented dashboard. Here's where it gets really interesting though. Yeah. We just laid out this strict theory of actionable inputs and keeping things to under 15 metrics, right? Yeah. Keep it tight. But if we look at the most rigorous, heavily studied, real-world implementation of this exact concept, which is Amazon under Jeff Bezos, there is a glaring contradiction. Yeah. The famous Amazon S-Team meetings. Right. Their senior leadership team. Our sources point out that Amazon famously used a 50 to 100 page metrics document for these weekly reviews. It's a massive document. Wait, 50 to 100 pages. That directly contradicts the 7 to 15 metric rule you just mentioned. I mean, how does a leadership team digest a 100 page document without getting entirely lost in the weeds? It seems like a massive contradiction on the surface, I know. But it is actually a masterclass in structural dashboard design. Okay. How so? Well, Amazon's S-Team document was indeed a massive comprehensive operating review. But the reason it functioned effectively, the reason it didn't just overwhelm Bezos and his executives, was because it was entirely built on a framework called driver trees. Driver trees. So instead of just a random list of 500 metrics, there's an actual hierarchy. Yes. Walk us through how a finance team actually builds a driver tree from scratch. It is entirely about building a causal chain. It's like a mathematical map of cause and effect. At the very top of the tree, you have your ultimate output metrics. Like total revenue. Exactly. Let's say total revenue or overall customer growth. But right beneath those outputs, the dashboard visually and mathematically descends into the input drivers that actually cause those outputs to happen. Oh, I see. So below total revenue, the tree splits into traffic and conversion rates. And then below conversion rates, it splits into page load speed or time to first purchase or defect rates in the fulfillment centers. So every number is essentially mathematically linked to the number above it. That's the key. So you aren't just looking at the weather outside. You're looking at the specific atmospheric pressure systems that are causing the weather to happen. That is the exact mechanism. And this is vital because it pushes all of the heavy analytical work into the design of the dashboard itself rather than saving it for the meeting. Which saves so much time. Right. Think about a normal corporate meeting. If revenue is down 8%, the leadership team will spend the next hour arguing about why. Oh yeah. Marketing blames sales. Sales blames product. Yeah, exactly. Is it a volume issue? Is it a pricing issue? But with a rigorously built driver tree, there's no argument. You just scan down the branches of the tree. You can just see the root cause. You can see instantly, "Ah, revenue is down because volume dropped 12% specifically in this one product because the page load speed on the mobile app increased by two seconds." Wow. You just saved an hour of political arguing and can jump straight into the decision, who is fixing the mobile app? And what I find brilliant is how Amazon forced this level of clarity organizationally. They had that famous no PowerPoint rule. Oh yes, Bezos hated PowerPoint. Completely banned slide decks in these major meetings. Instead, he forced executives to embed these driver tree metrics into six-page written narrative memos. Which is so much harder to write. It really is. When you think about it from a communication standpoint, it exposes everything. Bullet points on a slide presentation can so easily hide lazy logic. Oh, absolutely. You just wave your hands and move on. Exactly. You throw an impressive stat up there, talk over it, and click to the next slide before anyone can really question it. A written narrative forces you to actually prove in complete sentences exactly how your data supports your conclusion. You cannot hide a logical leap behind a bullet point. It removes the illusion of understanding. It forces the leadership to confront whether the input metrics they are looking at actually explain the reality of the output. If your narrative can't seamlessly connect the input to the output, well, your driver tree is broken. Now, Amazon shows us this incredibly brilliant way to structure a dashboard when an environment is relatively stable, or at least growing predictably. Sure. But what happens when the entire macroeconomic world changes overnight? Because relying on a static dashboard in a rapidly changing world is like trying to navigate a new city using a map from 1995. For physical roads have completely changed. Right. But you're still driving based on where they used to be. You're going to crash. If we connect this to the bigger picture, we really have to look at the ultimate modern case study of this exact phenomenon, which is Airbnb. Oh, this is a great example. Founded by Brian Chesky, Joe Gibbia, and Nathan Blachartzik. Prior to 2020, if you looked at Airbnb's original marketplace dashboard, the complexity was just staggering. Because they're a two-sided marketplace. Exactly. They had to balance supply and demand at a super granular city by city level. They had an entire funnel of metrics for hosts, and then a completely separate funnel of metrics for guests, and then all these quality metrics interlocking the two sides. It sounds like a machine. It was. The entire machine was optimized for massive global urban growth. Which makes total sense for the 2019 economy. But then, obviously, March 2020 happens. The world stops. The COVID-19 pandemic shuts down global travel entirely. Almost overnight, Airbnb's bookings dropped by roughly 80%. 80%. Yeah. That old dashboard, which was this absolute marvel of financial engineering optimized for global cross-border travel, was suddenly crashing. Every single growth metric was flashing red. The map from 1995 no longer matched the reality of the roads. So what does the leadership team even do in that room? If the driver tree is broken, do you just sit there and watch the numbers hit zero? This is where Airbnb's leadership demonstrated incredible operational agility. They realized their dashboard was measuring a business model that, temporarily at least, no longer existed. Right. So they completely redesigned their executive dashboard in real time. In the middle of the crisis. Exactly. They threw out the metrics that were optimized for urban tourism, and pivoted to entirely new metrics that reflected the new reality of human behavior. Like what? What did they start tracking? They started tracking average length of stay, because people were suddenly booking month-long rentals to work remotely. But that makes sense. They tracked the distance traveled by guess, because international flights were grounded, and everyone was driving to destinations within maybe 200 miles of their home. Yeah. The road trip boom. Right. And they started measuring non-urban host counts, because people desperately wanted to escape densely populated cities. That is such a vital takeaway for anyone building these systems. A dashboard is not a permanent artifact carved in stone that you just update month after month. No, not at all. It's a living instrument. It absolutely must evolve exactly as your business strategy evolves. If your strategy pivots to survive a crisis, but your finance team is still feeding you the old metrics, you aren't managing your business. You were just reminiscing about the past. You have to fly with the instruments that match the weather you are currently in. If the weather changes, you need different instruments. So Airbnb survived, because they aggressively aligned their metrics with operational reality. But what happens when leadership intentionally engineers a dashboard to ignore reality entirely?

[00:13:12:20 - 00:23:10:23]
The dark side. Yeah, because we've seen dashboards build empires and save companies in crises. But dashboards can absolutely be weaponized. They can be expertly crafted to lie to investors, to boards of directors, and to regulators. Which brings us to Theranos. Yes, Theranos. By 2014, the Theranos board was looking at a dashboard under Elizabeth Holmes that justified a peak valuation of $9 billion. $9 billion based on a lie. And this is all built on the promise of the Edison, right? That blood testing device supposedly capable of running hundreds of complex medical tests from a single pinprick of blood. Which we now know didn't work. So what does this all mean for the mechanics of our deep dive here? The Theranos board wasn't naive. It included incredibly intelligent people, former US secretaries of state, seasoned military generals, corporate titans. Oh, may smart people. How did a dashboard structurally manipulate them during their Monday morning meetings? Like, what were they actually looking at? They were looking at a masterfully curated fiction. And the how of that deception is so critical to understand. Because the board members were not biomedical experts, right? They relied entirely on the UI of the dashboard. They trusted the metrics they were showing. Exactly. Theranos leadership filled that dashboard with what looked like undeniable momentum projected revenue models, test volumes, and massive retail partnerships rolling out with companies like Walgreens and Safeway. So they just used these massive vanity metrics to paint a picture of inevitable scaling success? Yes. But the driver-tru was completely disconnected from the actual product. What was structurally omitted to hide the fraud? The operational truth. They aggressively siloed information so the board literally never saw the foundational input metrics. They actively hid the fact that the Edison device was failing. Unbelievable. They hid that the vast majority of their actual patient tests were secretly being run on conventional third-party commercial machines. They were just using normal lab equipment. Right. And most importantly, they completely excluded their internal quality control data from the executive dashboard data, which showed massive dangerous failure rates. The dashboard was completely decoupled from reality. It wasn't an instrument of truth. It was basically a marketing document disguised as an FP&A deliverable. And the fallout was catastrophic. We all know how it ended, leading the SEC to step in regarding the massive deception of investors while regulators at CMS, the Centers for Medicare and Medicaid Services, flagged the immense medical dangers. It ultimately resulted in a criminal conviction in 2022. A total collapse. But looking at the mechanics of that fraud, how does an honest company like a company that actually wants to know the truth avoid falling into the trap of just showing the board a rosy, curated narrative? It requires deep institutional courage. And our sources outline three specific structural rules to prevent this. Okay. What's the first one? First, you have to force uncomfortable metrics onto the primary dashboard. Things like product defect rates, severe customer turn, and employee turnover. So if a metric makes the executive team wince, it probably belongs on the main screen. Exactly. You can't hide from it. Second, you must use symmetric measurement. Symmetric measurement. How does that work in practice? It means you cannot only show the good news of a dataset. If your dashboard highlights new partnerships signed this quarter, you must right next to it, mathematically show partnerships exited or lost. Oh, that's smart. If you show new customers acquired, you show customers who canceled. You don't just show the incoming tide. You mandate visibility on the outgoing tide too. What's the third rule? Third, you have to empower an independent voice like an internal auditor or highly protected finance lead to look at the dashboard in a meeting and constantly ask, "What is missing here? What are we choosing not to look at?" You know, Theranos is a clear cut case of deliberate fraud. But sometimes a completely honest dashboard set up by well-meaning people accidentally creates an absolute monster. Oh, absolutely. And it happens simply because human beings are optimizing machines who will always respond to incentives. This raises an important question, perhaps the most critical question in organizational psychology. What happens when a mere measurement is transformed into a hard target? To understand that, we have to look at the Wells Fargo scandal. Such a classic case. For over a decade, Wells Fargo built a major piece of their executive dashboard around a metric called the cross-sell ratio. Basically, it measured how many individual products like a checking account, a credit card, a mortgage, an auto loan, a single customer held with the bank. Their target was eight products per customer. Right. They even coined a catchy internal slogan, "Eight is great." Which, on paper, tracking how many checking accounts or credit cards a customer holds sounds like a perfectly standard customer retention strategy. Yeah, makes sense. The logic is sound, right? A customer with five products is way less likely to leave the bank than a customer with just one. So where did the mechanism break down? It broke down because that metric didn't just stay as an observational tool on an executive dashboard. It cascaded down the organizational chart to the retail branch level. It became a weapon. It became the entire basis for employee compensation, for managerial promotions, and for brutal daily rolling quotas. Branch managers were pressuring employees multiple times a day to hit their cross-sell numbers. And if you didn't hit the target, you faced termination. Exactly. So, human behavior took over. What did the employees do? Finding the path of least resistance to survive over a decade, they opened roughly three and a half million unauthorized, completely fake accounts just to hit the numbers. Three and a half million. And this is the real tragedy of the dashboard. Up in the boardroom, the executives are looking at their screens thinking, "Wow, we are absolute geniuses. Look at our cross-sell ratio climbing year over year. We are crushing the market." Because the numbers were going up. Right. When in reality, their dashboard wasn't measuring success at all. It was fueling a massive systemic and toxic scandal. The fallout was historic. Wells Fargo paid $3 billion in fines and settlements. The CEO resigned. The Federal Reserve stepped in and slapped the bank with an unprecedented asset growth cap that choked their expansion all the way until 2025. This failure is the absolute textbook definition of Goodhart's law. Goodhart's law. Remind us what that is. Named after a British economist, Goodhart's law states that when a measure becomes a target, it ceases to be a good measure. Oh, wow. Yeah. The exact moment Wells Fargo tied employee survival and bonuses to the cross-sell ratio, the metric lost all of its informational value. Because humans optimized for survival, it was no longer measuring customer loyalty, it was measuring employee desperation. And it happened because the dashboard completely lacked paired countermetrics. If you were pushing a metric that hard, you have to balance it. Exactly. If you have a massive incentivized push for cross-selling, you must structurally pair it with a countermetric on the same dashboard. Like what? Like tracking customer complaints or measuring actual account utilization. An account with $0 in it and zero transactions is a massive red flag. And Wells Fargo didn't do that. No, they didn't. You have to constantly stress test your metrics by asking one incredibly cynical but absolutely necessary question. What is the absolute worst, most destructive way my team could achieve this number? If the answer involves fraud, harming the customer or destroying your brand reputation, you need a different metric. Exactly. So if we look at the common thread through all of these disasters and successes, you know, from Amazon's efficiency to Theranos's deception to Wells Fargo's toxic culture, it really comes down to fiercely protecting the operational truth of the data. That is the core theme of all of this. You protect the truth by ensuring every metric is strictly actionable. Right. You build driver trees so you understand the causal mechanism rather than just staring at a list of numbers. You evolve your dashboard the exact moment your macroeconomic strategy changes. Like Airbnb. Right. You pair your metrics symmetrically so you don't lose sight of the failures. And you constantly stress test your targets against Goodhart's law to prevent toxic incentives. I want to talk directly to you for a second, the listener. We have spent this entire deep dive looking at massive corporations. But remember the golden rule. What gets measured gets managed. Whether you are running a Fortune 500 company managing a five person startup or just organizing a community project, your dashboard isn't just a reporting deliverable. It is the absolute loudest voice in the room. It is actively shaping the attention, the priorities and the behavior of every single person who looks at it. And applying that logic actually leads to a rather provocative thought. Yeah. If we take the mechanics of the executive dashboard and apply them to our own personal lives, it gets quite uncomfortable. Okay, well then. Imagine if a totally independent objective auditor came into your life today and they built a driver based dashboard out of your last 168 hours, your last week on Earth. Oh, wow. That is terrifying. Right. If that auditor stripped away everything you say your priorities are, your stated strategy, and they only looked at the hard measurable data, your actual screen time down to the minute, your bank statement line items, your calendar blocks, what would that dashboard say your actual life strategy is right now? What are you actually optimizing for? And more importantly, does that hard data match the narrative you tell yourself? That is a staggering question to end on because the numbers ultimately do not care about our intentions. They just show us the reality of our actions. Next time you walk into that Monday morning meeting or even just look at your own daily calendar, remember the numbers you choose to focus on are quietly ruling your life. Make sure they are the right ones. Thank you so much for joining us on this deep dive. We'll see you next time.