A to Zipf [Podcast]

Welcome to another episode of “The Purposeful Banker,” a podcast brought to you by PrecisionLender, where we discuss the big topics on the minds of today’s best bankers. I’m Dallas Wells, and here with me is our cohost, Jessica Stone.

Hey, thanks for joining us. Today we are talking about mathematical principles, which sounds so exciting, I’m sure, but we’re going to talk about how those can come into play at a bank and then what you can do with this knowledge, so today’s topic is inspired by Dallas’ blog post from the other week, Zipf’s Law and Your Lenders, and I’m just going to put it out there in advance, I’m probably saying that wrong. I’m probably going to mess it up a couple times, so Dallas, feel free to correct me.

Yeah, those P and then an F things are a little tricky for us English speakers, so yeah, good luck with it.

Awesome. As I mentioned, Dallas, the mathematical principle we’re talking about is Zipf’s law. Can you explain what that is?

Zipf’s law is actually a mathematical principle that comes from a linguist at Harvard, of all people. The background is actually sort of interesting of where it came from. George Zipf was a linguist at Harvard in the mid-’30s, and what he started doing was what, apparently, was a common thing at the time for a linguist, is they were studying the frequency of word usage, so they would tally up in all the different languages what word was most frequent, second frequent, so on, and they would take this out for several thousand words, and some interesting things started coming out of that. No matter the language, it was like this universal rule that it was a power law distribution.

Specifically, the frequency of a word was inverse to its rank. You would have the most common word, which in English is the word “the,” and then the second most common word would show up half as many times, and the third most common words would show up a third as many times as the word “the,” and so it goes all the way on down until eventually the hundredth word shows up 1/100 as many times as the most common.

The interesting thing about this was that Zipf had no idea why this was the case and wasn’t the first one to come across this distribution showing up, and we still don’t know what the cause of it is. As they were digging deeper into it, they found that not only was it universal among languages, but it was in subsets of things too, so for instance, in individual books, so you could take “Tom Sawyer” or “Huckleberry Finn,” you could take the King James Bible. The frequency tables would be different, obviously, the Bible versus a Mark Twain book, they’re going to have different sets of words in there, but the distribution stayed intact, and they found this from ancient tablets that we still can’t translate. The characters were still coming out in that same distribution and it makes this nice power law curve when you chart it out, so we put a couple of those curves in the blog post.

This is where it started to get really interesting, though. Zipf found this over and over again in language, so he started looking at other things too. It also applied to things like city populations. The most populous city in the United States would be twice as big as number two, three times as big as the third one, et cetera, and not only was that the case, but that distribution stayed intact over time. Ten years later, the population of those cities would all change, but the distribution remained intact. The biggest city grew, the second city was still half that size. They looked at other countries, since kind of thing.

Now we find the same distribution in really random things, protein frequencies in genome sequences, things like Twitter followers of NBA teams, Facebook likes of NBA teams. They all keep following this distribution, and so it was just this crazy, almost seems like a coincidence, but it just shows up so many times. The basic thought is that the human brain is just kind of lazy or always looking for efficiency, so the most common word we use, we use those most common words all the time in place of other ones because they’re easy, they’re easily recognized for everybody we’re talking to, and so you pretty quickly get to where most of our conversations are made up of 50 words or so.

That sounds a lot like the Pareto principle or the 80/20 rule. Is that related, or am I just making that leap?

The 80/20 rule, the more shorthand one, states that for many events 80% of the effects come from 20% of the causes, or, to put it more plainly, 80% of your results will come from 20% of your efforts, so it’s a nice shorthand version of the more complex math behind it, of really trying to get people to focus on what are the things to do that are really going to drive results.

Pareto’s principle, as you dig into that one, Pareto was another guy with maybe too much time on his hands where he’s finding these 80/20 rules all over the place. He finds them in things like out in his garden where 20% of the pea plants produce 80% of the peas. I don’t know how you measure such a thing, but he took the time to do it and it’s, again, one of those things that shows up over and over and over.

Dallas, let’s bring this back to banking. How does Zipf’s law and that 80/20 notion relate back to banking?

Yeah, it comes back in lots of ways, just like that principal shows up over and over again all over the place in social things and in nature, it shows up in a lot of places in banks, too. A couple big ones worth paying attention to is your customer profitability. We didn’t do deep analysis on this, frankly, because there’s just lots of data and I didn’t take the time to go into it all yet, but I just pulled a handful of our client banks and looked at their most profitable customers, and you can just kind of eyeball, go down about 20 down the list, and it’s another Zipfian distribution.

The one we really paid attention to, though, is a topic that we’ve covered a couple times this year, we covered in the book that we’re releasing, the top lenders in every bank and how important they are and what that really means for managing those lenders and for managing your balance sheet. We dig into banks and their portfolios and are we getting the right results from the pricing and from the lenders and from the processes we have in place, and guess what, the old 80/20 rule shows up, where 80% or more of the results are coming from 20% of your lenders.

What effect would that have for the bank? If you see that, you say, “Okay, well, so what?”

Here’s the difficult part for banks, is just the nature of the business. Bankers worry a lot about consistency and about those one-off events that can be, frankly, dangerous for banks, so you make a thousand loans, it only takes one being an epic fail for you to derail the other 999 that you did. What that means is that bankers look for anomalies and they look for outliers and they try to fix those.

We’re all about a low-margin business that we make up for with scale, so lots of zeros on the dollars, and we make a little bitty sliver of a penny on each thing that we do. We just do it over and over and over again, so there’s no room for those kind of outliers, so what ends up happening is, as an example, our system, so PrecisionLender we put into priced commercial loans, and we put that thing in place and bankers say, well, I want all my lenders to use it exactly the same way, I want consistency, I don’t want any outliers, I don’t want any mistakes. When really, you’ve got 20% of your lenders doing 80% of the volume, and really, that’s probably understating the effect a little bit.

Banks are putting time and attention and money and building all these rules and systems into really corralling or controlling some of the bank’s least productive lenders. We’ve got 20% of the results coming out of 80% of the head count, and we’re trying to keep them from making mistakes. Carl, our CEO, was the one who, of course, noticed – he’s the resident math nerd – he noticed the Zipfian distribution when we charted these things out, and of course once we saw that we start seeing it all over the place. We see it in our own client base over and over in terms of size and revenue to us, in terms of usage of our tool, just over and over again. That same curve, we see all over the place now.

Carl describes this curve for bankers like this. If a bank’s top lender generates a dollar of returns – however you want to measure that, it could be a dollar of portfolio size, it could be a dollar of risk-adjusted net income, but again, a dollar of net value to you – then the fourth best lender would generate a quarter, and the hundredth best lender would generate a penny, so as shorthand, that’s what he calls them. We’ve got dollar lenders, quarter lenders, nickel lenders, on out, and eventually you get all the way to the penny lenders, so if you think about what the banks are doing system-wise, they’re putting rules in place to keep their penny lenders from making any mistakes.

They’ve got Billy Bob out in the Timbuktu branch and they want to make sure that he doesn’t put deals into the model wrong, so they put really restrictive rules in place, and there’s lots of penny lenders, just by the nature of the distribution. There’s a ton of them, and so we spent all of our time chasing them around and following up and saying, “Hey, Billy Bob in Timbuktu, you did this wrong again, so let’s just turn off that functionality of the tool,” or “let’s put a policy in place where you’re not allowed to do that,” and so we’re spending all the time and effort corralling those things.

In some ways, that’s a meaningful thing, right? We don’t want bad loans put on the books by those penny lenders, so I don’t want to imply that we’re overlooking credit losses. Credit losses is its own animal, and those, guess what? Follow the same distribution. We don’t want to make any big mistakes there. What we’re talking about is pricing, giving your lenders the ability to put deals on the books and to not make a dumb pricing decision.

The frustration comes at the other end of the curve. All those rules and restrictions and policies and things that you put in place, they’re slowing down your dollar lenders, and they’re the ones that are dealing with your best clients in mass volume. They’re the ones allocating all the bank’s capital and the loan portfolio and really generating the returns. For those lenders and for their customers, specifically, the bank, philosophically, would be willing to bend the rules.

We’d be willing to be really responsive, and maybe as a rule we don’t do 10-year fixed-rate real estate loans, but for the number two customer in the bank who has tons of other business with you, would you maybe do a 10-year fixed-rate deal if that made him happy? Yeah, probably so, but we’ve put all these rules in place, so now that lender has to go and get special exception approval and you’ve restricted it down in the system where they can’t even get it put in there, and there’s all these extra steps and all this extra frustration that gets put in place.

The impact is not way out on the curve for the penny lenders. The impact is for your top four folks that are generating all the business with your very best customers.

Dallas, we’ve actually seen this play out. In fact, you and I heard this exact thing from a client recently. Want to explain that situation and what they were facing?

This was basically that same scenario, and this was a bank that had done an acquisition and so there were some new folks in and they’re trying to get consistent and disciplined in how they do things, so they’ve got two different cultures they’re squishing together, and that’s part of the issue here, but they had a leadership group that was placing pretty much all the emphasis on net interest margin, and it’s a good, solid earning bank.

They’re really good, but the management says, “Hey, margins are compressing and we don’t like how that’s coming about. It feels like it’s from sloppy pricing, so we’re going to put some restrictions on, don’t do any loans below a certain rate,” and that’s an issue we talked about a lot. “Don’t do any fixed rates beyond this,” and “we want spreads of at least X on any floating-rate loans,” so what they’re trying to do is establish that consistency.

They feel like they’ve got, in the scattered markets that they’re now trying to corral, they’ve got people doing things that are outside of the comfort zone. We look at their distribution of production and they had four lenders putting almost all the loan volume on the books, and those folks are getting really frustrated with this. Well, if I have to do a deal that starts with a five on the interest rate, if I do a fixed-rate deal, I can’t service my very best customers. They’re going to walk, so yeah, the hundred loans we did last month, 98 of them will be above there and we’ll be able to high-five each other and say “great, we held onto margins,” but we lost the deal, the one deal that, by volume, by size, by actual dollars of income, would have dwarfed all those other ones.

It’s a hard thing to measure because it’s lost deals. There’s frustration with all the extra rules and restrictions, and that’s what we were hearing from this bank, was, how do we deal with that? How do we keep from giving up margin, but do it in a way where it doesn’t just make it hard to get deals done and to do business?

What was your suggestion for the bank on how they might address this situation?

Yeah, so I feel the pain, so to speak, of the management team, and I know where that’s coming from. I’ve been a part of banks facing that same kind of margin compression, and that is the reaction, is say, “Hey, guys, knock it off. Quit doing deals that are going to squeeze us like this.” I think the only way to look at it is, you can look at the simple algebra of it and there are some deals that you can put on that will be below those spreads that you’re after, but just the straight dollars of income and the fact that you’re allocating dollars, maybe, away from something like the bond portfolio and into a loan, you can hang onto margins while still doing some skinnier deals, but what you have to show is what that volume and trade-off is, rate volume trade-off.

That’s the one we talk about all the time with banks, is if you’re going to cut rates by this amount, how much more volume do you have to do to make up for it? We’ve got to try to quantify that somehow, and again, we’re talking about missed deals, so we do have in PrecisionLender a way to capture competitive offers. The problem is, the banks have to be pretty disciplined about putting those in there so that we can get enough data points to go off of, so that’s always the first stop is to go look at those deals that they’ve marked as lost to competition. We can see what was the offer that they lost to, and we can do some real analysis with that.

When you don’t have that, then what we have to do is more of a hypothetical. Luckily, since we’re talking about an 80/20 rule, or in this bank’s case we’re talking about four lenders, is let sit down with those four lenders and let’s say, “All right, guys. You feel like there’s some frustration, a lack of ability to be responsive to your clients, and you feel like there’s maybe some deals that we lose out on. Let’s do some hypotheticals here. Here’s where you’ve been asked to price it. If you could price it where you think it should be, an eighth or a quarter point lower, how many of those deals would you have won? Let’s write down some specifics,” so have them just pencil down, well, here’s four deals that we lost last quarter that we feel like we would have gotten.

Okay, plug those in. Instead of that money sitting in the bond portfolio or in fed funds sold, plug it into the loan portfolio with those skinnier spreads and let’s see, well, what would margins have been with and without? You’ve got to quantify it for that management team so that they get a feel for those few deals, for those very best customers, and really, what we’re talking about is for those top lenders, is there a way that we can give them the tools to make smart business decisions?

We’ll trade a slightly skinny deal to a very solid customer rather than have it sitting in a one-year agency bullet earning effectively nothing. This is how we try to approach it with the management team, is put a few numbers to it and show them that what we’re doing is, instead of … If we want to lift the performance of the bank, there are a couple ways to go about it, so if we want to add a nickel of performance, you can go down the curve to all your nickel lenders and you can say, “All right, nickel lenders. We want you to be twice as good,” or you can go to your dollar lender and say, “I want to make you a dollar and five cent producer,” so which one do you think is more likely to actually happen, and where does it take the least amount of effort to get the most amount of results?

Also, when you’re talking about a bank wanting to really move the needle and really get better performance or change the asset mix or adjust the credit quality, where you’re most likely to make a dent and be able to change those things is by the people putting the most volume on, so let them be flexible.

Communicate with them on what the real strategy of the bank is and what you’re trying to achieve, and give them, basically, a different set of rules than everybody else on the curve. You let those alpha lenders, those ones that are really good, let them do what they do best, and then secondarily worry about penny lenders and penny mistakes. It’s not something we want to ignore, but that’s not where all your time and attention should be. Focus where the results are.

Dallas, any other examples that come to mind of maybe this kind of exceptions for those dollar lenders come to mind?

Yeah, so that’s the easy one, is pricing exceptions and getting those folks a little more leeway. It comes down to even just basic management style, so an example came to mind from one of the banks I was at where we had … It was a smaller bank. I think there were five people doing commercial loans, and there was one guy who did 90% of the volume, so he was a big producer and dealt with all of our top clients, so to try to get the other four to measure up, the CEO was putting call programs in place and saying, “Well, I want you to do this many dials and this many cold calls and visit this many customers in a week, and you’ve got to put it all into this spreadsheet,” this homemade hack of a CRM system, “and I’m going to track these, and this is also what I’m tying your bonus too.”

That may have been somewhat helpful to the four who were doing essentially no volume, but it was a giant pain in the rear for the guy who was really busy because he’s doing 90% of the commercial loan volume, so he was being held to that same standard when he didn’t need the help, he didn’t need the tracking system, he didn’t need to be told how many cold calls to make. He’d clearly figured that out. Guess what? He got frustrated and he started looking for a new place to take those loans. He wanted to work somewhere else where there wasn’t in all those rules in place, so it really comes down to trust.

Once you have somebody who’s that kind of a performer, you’re going to be paying them a ton of money because that’s the going rate for somebody who’s putting that kind of volume and that kind of revenue into the bank, trust them as professionals, as big contributors of the bank, as meaningful members of the team. Bring them into the discussions on where the bank’s headed as a strategy, let them know what we really need from the loan portfolio, and they have a different set of expectations and rules. Then, you’ve got the machine, so to speak, the rest of the lenders.

Let’s be consistent with those and let’s put some rules and restrictions in place so that we get at least disciplined and consistent decisions and production out of them. For those top few at the very top of that Zipfian curve, leave them be. Enable them. Give them some tools that will help them be flexible and responsive to the customers rather than trying to put rules in place that restrict them, and that’s probably the big takeaway, is enable them rather than try to box in everybody else.

Awesome. Thanks, Dallas. I think that will do it for us for this episode, so thanks everyone for listening. We will provide some links to Dallas’ blog post and maybe a couple resources in the show notes for this episode. You can always find those at PrecisionLender.com/podcast. If you like what you’ve been hearing, please make sure to subscribe to our feeds in iTunes, SoundCloud, or Stitcher, and we love to get ratings and feedback, too. Thanks for tuning in. Until next time. This has been Jessica Stone and Dallas Wells, and you’ve been listening to “The Purposeful Banker.”