Chapter 4

Becoming the Expert

Minute Fifty-Two

It's the fourth webinar this week, and by minute fifty-two nearly everyone has left.

The attendee count started at two hundred and eleven. He's been watching it fall. At minute forty, when the slides ended and the Q&A began, it dropped by half. It's minute fifty-two now and the counter says nine, and he suspects seven of those are tabs somebody forgot to close.

A month ago he'd have been one of them. Industry webinars were background noise: necessary for staying current, not worth a whole hour. He'd half-listen, answer email, and pull one statistic into a pitch deck if something sounded quotable.

Now there's a notebook open next to the laptop and he's filled four pages by hand.

The presenter is a consultant from one of the big advisory firms. Most of what she showed was familiar: delinquency rates, refinance projections, the usual quarterly numbers. Then, near the end of the slides, she put up one that made him stop writing entirely.

Sixty percent of borrowers cannot recall their loan officer's name two years after closing.¹

He'd seen it before. Read it in some report months ago, filed it under interesting, moved on.

Each of those borrowers cost the lender thousands of dollars to acquire, and most of them can't name the person who walked them through the biggest financial decision of their lives.

He opens the file Tom told him to start. It began in a parking lot as What I Know About Client Retention in Mortgage; now it records what he notices. He calls it the spark file because "the document" sounded odd out loud.

Entry 34: The Name Recognition Gap Lenders pay for the relationship and then let it go. The refinance never comes, and neither do referrals. What would it actually take to be memorable? Source: webinar slide, figure unverified.

The Q&A turned out to be the part worth staying for. A woman from a credit union in Ohio asked how you'd measure retention at all if the loan officer who originated it has since left the company. The presenter didn't have a good answer. She said something about attribution windows and moved on quickly.

Marcus wrote the question down and put an asterisk next to it.

It's ten past six in the evening. Nobody is paying him to watch a Q&A on borrower attribution, and there's no exam.

Ask yourself again after Rachel decides, whichever way it goes, Tom had said. Rachel decides in a little over a month whether to keep funding the company, so Marcus can't tell yet whether this is fascination or panic, and that irritates him.

He asks himself whether he's fascinated, and the word doesn't fit. Fascinated people collect things for fun. He's doing it with a deadline and a spreadsheet.

Then he plays back the last four minutes of the recording to get the Ohio woman's question down exactly, and he notices himself doing it, and he stops arguing.

The Gurley hardcover is on the desk with the receipt still in it. He hasn't finished it. He has read the same eleven pages four times.

The Thing That Isn't in Any Report

The next afternoon he pulls the usage data from VaultPath's five beta lenders.

Small sample. Real data. He's looked at it a dozen times for product reasons: which features get used, where people drop off, what to build next. He's never once looked at it for what it says about borrowers.

Today he does.

Three of the five have the automated check-ins switched on: the birthday notes, the home anniversary messages, the credit-score updates. Two have them off. He expects a clean split, and for about ten minutes he has one. The "on" group replies at a clearly higher rate than the "off" group. Not opens. Replies. Borrowers writing back eighteen months after closing, asking things, starting conversations.

He's most of the way through drafting that finding when he notices the third lender.

They have the feature on. They're performing worse than both of the lenders who have it off.

He goes looking for the difference and finds it in twenty minutes. The two lenders who outperform send fewer messages. Each is tied to something true about that borrower: the anniversary of their closing, or a rate move that affects their loan. The third sends the same note to everyone on the fifteenth of the month.

So automation isn't the cause. The lender that lost used it too.

The all-three number was dragged down by the bad lender. Take it out and the other two reply at roughly three times the rate of the two with the feature off, but that rests on two companies instead of three.

Entry 35: The Response Pattern Fewer, specific check-ins: ~3x the replies of the two lenders with the feature off (n=2, be honest about this). One lender sending the same generic note on a fixed monthly date did worse than the lenders with it off (n=1). Source: VaultPath usage data, five beta lenders.

He puts five asterisks on it.

Then he does something he hasn't done since the cold-email disaster in January. He opens the machine and asks it what drives borrower re-engagement in mortgage servicing.

The answer arrives in about four seconds and it's good. It's organized and clear, with citations that look real, and it covers personalization, lifecycle marketing and touchpoint cadence. It's better written than anything he heard in four webinars this week. It's the average of what's been published on the subject.

It doesn't contain his pattern.

The model has never seen his lenders' usage logs, and nobody has published anything about them.

He reads the answer twice and adds a line to Entry 35: Write this up.

Thirty-Eight Entries

By the end of the week there are thirty-eight entries.

Some are statistics from webinars. Some are things a beta customer said offhand on a support call. Some are questions he can't answer. Those are turning out to be the most useful entries, and a month ago he would have thrown them away.

He has thirty-eight loose notes and no routine for turning them into a post a stranger would read. Rachel decides in a little over a month.

The Expertise Learning System

You get good faster with a routine than by waiting until something catches your interest.

Five steps: Discover, Capture, Refine, Publish, Analyze. A machine can now take real work off every one of them, but not the judgment in each.

Discover. Choose what you read instead of taking whatever the feed serves up. Thirty minutes a day on curated industry material. An hour a week on something longer and harder, such as a report or a webinar you actually stay in. Ninety minutes a month on primary research, meaning conversations with people. Nobody else can copy that input.

The machine's half: reading at volume. It'll read the hundred-page report and hand back the twelve pages that matter, summarize the earnings call, compare the regulation against last year's version, and keep you current on a subject that would otherwise eat a day a week.

Yours: noticing what makes you stop. The machine can tell you what a hundred reports say. It can't tell you which one bothered you, and that is where your next question comes from.

Capture. Keep everything in one place. A system split across three apps gets abandoned inside a month. Each entry should have the observation, where it came from, what it implies, and what it makes you wonder. Call it a spark file or anything else, as long as you can find an entry eight weeks later.

The machine's half: the clerical work. Transcribe the customer call. Tag the entry. Surface the four things you wrote six weeks ago that touch what you're writing now. Human memory is bad at that, and it's why most people's notes never get used again.

Yours: deciding what earns an entry, and marking the ones you'd argue about. A system that keeps everything fills up with entries you'll never read.

Refine. Spend thirty minutes a week reviewing your entries. Look for the same observation said different ways and, above all, for contradictions: places where your own evidence disagrees with what everyone in your industry knows.

Refine is where the machine helps most and misleads most. It's good at holding forty entries at once and telling you which ones go together. It also leans toward consensus, because it's built from what has already been published. It will find the pattern that's in the literature. If you let it referee, it will rank the pattern that's only in your data below the one everyone already knows, and it will give good reasons.

So run the contrarian test yourself, against your own evidence rather than your instincts: Where does what I've actually seen contradict what everyone says? Where has something worked that best practice says shouldn't? Where has a best practice failed in front of you more than once, in a way you could defend? Those entries are the ones worth building on, and the ones a model is most likely to rank lowest.

Publish. If it stays in your head, nobody else benefits and you never find out whether you're right. Publishing tests the thinking and lets strangers find you. An observation and a question is enough to start.

The machine is a competent editor. It will tighten a paragraph and tell you the third sentence is doing nothing. The rest is yours, and it is more than you'd like. A finding drawn from data only you can see will name somebody, and sometimes that somebody is a customer. Expect to hesitate most on your best findings.

Analyze. Watch what happens, and watch the kind of thing that happens. Likes tell you little. Comments tell you more. A message saying this is exactly the problem we have is the strongest signal, because it comes from someone who has the problem, and it feeds straight back into Discover: now you know which of your questions is also somebody else's.

The machine's half: the counting, and the pattern across months you'd never hold in your head. Which topics, which formats, which openings.

Yours: judging which numbers to trust. A post can do well for reasons that have nothing to do with whether it was right, and a model told to raise engagement will keep nudging you toward the average post, one reasonable suggestion at a time.

This week, open one file and write down the next thing that makes you stop. That's your first entry.

The One He Doesn't Send

Ten days later there are forty-seven entries and nineteen asterisks. One of the asterisked entries has grown into a draft post.

It's the automation finding, and it has grown from an observation into an argument about how lenders do retention.

He writes it standing up, in one pass, which he's never done.

Most lenders handle retention as a marketing task, and the money they spend on it isn't working.

A lender spends roughly $3,700 acquiring a customer through paid channels, close to $2,000 organically.² Then, at the moment that relationship is worth the most, it stops maintaining it. By the figure everyone quotes, six in ten of those borrowers can't name the person who served them two years later.

The usual answer is more automation. I've been looking at data on this that I can't show you, and I don't think that answer holds. The lenders sending fewer, specific messages get about three times the replies. The one sending the same note to everyone on the fifteenth of the month did worse than the lenders who have it switched off.

The difference is whether the note is about the borrower. Some lenders are spending real money to do worse than the ones who skip it.

And the economics of getting it right aren't marginal. The study everyone still quotes found that cutting customer defections by five percent raised profit by twenty-five to ninety-five percent, depending on the business.³ That paper came out in 1990, and we're all still citing it.

He reads it back.

It's clumsy. The second paragraph is mostly numbers, the third hedges in the middle and then stops hedging, and the whole thing reads like a man reading his own notes aloud. He can hear that it isn't good. He can also hear that it's true and that it's his.

Most of what he has written in eighteen months was smooth and said nothing new.

He doesn't post it.

He closes the laptop, opens it again, reads it a third time, and still doesn't. Then he types one more line underneath, for himself, the way he's started doing.

I'm afraid three of my five customers can read this and will know it's about them.

His phone buzzes at 9:40. Tom.

"How's the capture going?"

Marcus looks at the draft sitting unsent on the screen. It's written badly and he hasn't shown it to anyone.

He types: "I found something. I don't think anyone else can see it. I've had it written for an hour and I can't make myself publish it. It's also not very good."

The reply takes under a minute.

"It's probably better than you think."

Then, a few seconds later:

"Send it to me if you want. That doesn't count as posting it."

Marcus doesn't post it that night either.

He also writes to Rachel, who asks about the sales calls less often now.

"Can we do Friday? I have something to show you. It isn't revenue yet."

Endnotes

The 60% figure is industry lore. It circulates constantly in mortgage conference decks, and I have not been able to trace it to a primary source. I have left it in the book anyway, and readers deserve to know that. After thirty years in this business my judgment is that it understates the problem rather than overstating it, but that is my judgment, not a measurement, and it should be read that way.

What is measurable runs the same direction. STRATMOR Group reports that only 18% of borrowers return to their servicer when it is time to originate again (Mike Seminari, "Strangers No More: Flipping the Script on Borrower Retention," May 2025, https://www.stratmorgroup.com/strangers-no-more-flipping-the-script-on-borrower-retention/).

Whatever the precise recall number is, the relationship decays unless somebody deliberately maintains it. That is the argument for the polite, persistent re-engagement system this book goes on to build.
Focus Digital, "Average Banking Customer Acquisition Cost," August 2024, https://focus-digital.co/average-banking-customer-acquisition-cost/, reporting mortgage acquisition costs averaging $3,664 through paid advertising and $1,963 organically. Marcus's "roughly $3,700" and "close to $2,000" round these. Readers should weigh the source for what it is, an agency's aggregation of industry figures rather than primary research, and treat the numbers as the right order of magnitude rather than a measurement.
Frederick F. Reichheld and W. Earl Sasser Jr., "Zero Defections: Quality Comes to Services," Harvard Business Review, September–October 1990, 105–111. Working across a set of service industries, the authors found that reducing customer defections by five percent raised profits by 25 to 95 percent, with the size of the effect varying widely by industry. Financial services appear at both ends of that range in their own data, so the high number should not be read as the expected one. The paper is still widely cited.