The Mortgage KPIs I Actually Watch (And the Ones Most Lenders Get Wrong)

I've spent 30+ years inside mortgage lending operations. I was Employee #7 at DeepGreen Bank in 2000. I built EquityOnline as VP of National Home Equity at Quicken Loans. I founded Kaleidico in 2005 and have run lead generation and marketing engagements for a hundred-plus mortgage lenders since. I've watched the industry through the 2008 crash, the refi booms, the QM era, the rate spikes, and now the AI-plus-compliance era.

Across all of it, the most consistent pattern I see is this: most mortgage lenders are measuring the wrong things. They track metrics that feel important because they've always been tracked, while the metrics that actually determine whether the business makes money go unmeasured or get measured so poorly that the data is meaningless.

This post is the short version of what I watch when I walk into a lender's operation, and what I've learned to stop paying attention to.

The KPIs I watch

Here are the metrics that actually predict whether a mortgage lending operation is going to be profitable in the next 12 months.

Effective CPA per funded loan, by channel. Not cost per lead. Not cost per application. Cost per funded loan, broken out by every acquisition channel. This is the number that tells you whether your marketing spend is actually producing revenue, and it's the number most lenders either don't calculate or calculate so loosely that the channels can't be compared. When I'm auditing an operation, this is the first view I ask for. If the lender can't produce it in a week, the operation has a data problem that's almost certainly costing them money.

Speed-to-first-contact. The data has been unchanged for 20 years: a lead contacted in the first five minutes converts at a multiple of the same lead contacted at 30 minutes, which converts at a multiple of the same lead contacted at two hours. I built icoSales at Kaleidico specifically to attack this metric. If a lender's median speed-to-first-contact is measured in hours, their sales operation is leaving conversions on the floor every day.

Contact rate per lead source. What percentage of the leads I bought from Vendor X did my team actually get on the phone? This is the metric that separates good vendors from bad vendors, and it's almost entirely independent of lead price. A cheap vendor with a 20% contact rate is more expensive than an expensive vendor with a 45% contact rate. The effective economics don't show up until you measure this.

Application-to-funded rate, by source cohort. A 60% application-to-close rate on pre-qualified leads is a completely different number than a 60% rate on unqualified leads. Aggregate conversion numbers hide the variance that matters. I insist on seeing this metric cohorted by source and product mix, because the averages lie.

Dial attempts per lead. Most mortgage sales floors under-dial. The research is consistent across verticals: 8-12 attempts over 14 days is a reasonable baseline. Most floors average 2-4 attempts before abandoning a lead. The lenders that enforce a disciplined cadence convert materially more loans from the same lead inventory. This metric surfaces the discipline problem.

Lead-to-funded payback period. How many weeks or months between when the lead cost was spent and when the funded loan revenue hits the P&L? This matters for cash flow planning and for identifying channels where the economics are structurally broken even if the nominal CAC looks acceptable.

Loan officer productivity, measured in funded loans per week. Not calls made. Not applications taken. Funded loans per LO per week, with a rolling four-week average. This is the only LO metric that matters at the operational level, and it's the one I've watched lenders under-measure the most.

The KPIs I've learned to ignore

Here are the metrics that get obsessive attention in most mortgage operations but tell you almost nothing useful.

Monthly funded volume in absolute terms. Celebrating a record month is fine, but absolute volume without unit economics context tells you nothing about the health of the business. I've watched lenders hit record volume while burning cash on negative-margin loans. The volume number is a vanity metric unless it's paired with a margin view.

Raw application counts. An application is an ambiguous object. Some lenders count every form submission as an application. Others require full documentation before the count increments. Comparing application counts across lenders, or even across periods at the same lender if the definition drifted, is essentially meaningless.

Average loan size. Optimizing for average loan size usually means chasing jumbo loans at the expense of diversified business. The metric drives portfolio concentration, which is the opposite of what you want during a rate cycle. I track loan size distribution, not averages.

Raw lead counts. The number of leads bought this month tells me nothing about whether the marketing spend is productive. A lender buying 5,000 leads at a $40 CPL with a 5% contact rate is often running a worse program than a lender buying 1,500 leads at $80 CPL with a 42% contact rate. Raw counts without quality context are noise.

Pull-through rate as a headline number. I care about pull-through at the cohort level, not at the aggregate level. Aggregate pull-through hides the source-level variance that actually tells you what's working. A 72% aggregate pull-through that's a 90% rate on your best source and a 35% rate on your worst is a different operational picture than a uniformly mediocre 72%.

Why the industry gets this wrong

Three reasons the wrong KPIs dominate mortgage lending dashboards.

First, path dependency. The metrics most mortgage operations track today are the metrics they tracked in 2005, with minor iterations. The industry has gotten dramatically more sophisticated in data infrastructure, but the KPIs that get executive attention are still mostly inherited from an earlier era.

Second, LOS and CRM default reports. Most lenders live inside their loan origination system and their CRM, and the default reports those systems generate happen to emphasize the legacy KPIs. Changing what gets reported requires effort that most operations teams don't have the bandwidth to invest, so the default wins.

Third, executive discomfort with unit economics. Lenders that started in high-volume, low-competition markets could cover operational waste with transaction volume. The executives who ran those operations are not natural unit-economics thinkers. Moving to a margin-first, effective-CPA view feels like a cultural shift, not just a reporting change.

None of these are good reasons to keep measuring the wrong things. They're just the reasons the wrong measurement persists.

What to do about it

If you're running a mortgage lending operation and you want to stop measuring the wrong things, the fix is neither fast nor glamorous.

Start with one week of building a one-page view that shows effective CPA per funded loan, contact rate, and application-to-funded rate, broken out by every acquisition channel. If your team can't produce this view in a week, you have a data pipeline problem that needs to be solved before any other KPI work matters.

Once you have that view, run it weekly. Make it the opening dashboard in every operations meeting. Kill channels where the effective CPA exceeds your margin envelope. Double-down on channels where the effective CPA is below your target. Ignore vendor claims that contradict your own cohort data.

Then add the cadence and speed-to-contact discipline. Most operations gain 15-25% in funded loan throughput just from tightening cadence and first-touch speed, without any additional marketing spend. That's pure margin.

The lenders I've worked with who committed to this kind of KPI discipline tend to outperform their peers not by some flashy marketing advantage but by the compounding effect of measuring and optimizing the right things over a multi-year period. The lenders who keep tracking monthly funded volume and average loan size keep being surprised when their margin compresses.

The bottom line

Mortgage lending is not going to get easier in the next five years. Rate volatility, compliance complexity, AI-driven competitor efficiency, and the ongoing consolidation wave are all going to pressure unit economics. The lenders that survive and compound will be the ones who measure the right things and make operational decisions based on what the data actually says.

I advise mortgage lenders on exactly these questions through Kaleidico and Bill Rice Strategy Group. If your operation is running on the wrong KPIs and you want to fix it, reach out.

And if you want the full lead-buying side of this discipline, that's the subject of my book,

The Lead Buyer's Playbook. Free online or on Amazon.