Sore Thumb Resource pack Run the numbers
§6

Results

What our own accounts actually produced. Insurance, where the system was built, and mortgage, where we are newer. The weak accounts are left in.

Confidentiality and method

How to read every number on this page

Why no company namesClient agreements carry confidentiality terms, so accounts are anonymised. On a call we open the live dashboards and the ad accounts behind them.
What an enquiry isA person who finished the funnel or the form and consented to contact, counted in that account's own ad reporting over the stated window.
What we do not countClicks, impressions, video views, and anything Meta's lead column re attributes between ad sets. Funded loans are counted from the client's CRM, never from Meta.
What this is notNot a projection, not a typical result, and not a claim about what your brokerage will do. Some of these accounts worked and some did not.
Mortgage

The one brokerage month we can show

One mortgage brokerage, month three of the engagement. This is the only funded loan data we hold in the vertical, and we say that out loud rather than dress it up.

68 enquiries

In the month, from that brokerage's own ads and funnel.

9 funded files

Counted in their CRM, in month three of the engagement.

$65 to $24 per enquiry

Cost per enquiry at the start of the engagement against cost per enquiry by month three, same account.

What we will not do with thisWe will not multiply it out across your officer count and call it a forecast. One account, one month, named on a call if that brokerage agrees to take it.
Our own account

We run the same system on ourselves

Sore Thumb sells to mortgage brokerages using the exact system in this pack: Meta ads, a qualification funnel, instant follow up, calls booked into a calendar. These are our own numbers, from 17 September to 6 October 2026.

LineFigureSource
Ad spend$3,236Our own Meta account
Applications completed43Our funnel, deduped by email
Cost per application$75Spend divided by applications
Calls booked and held15Our calendar, matched by email
Cost per held call$216Spend divided by held calls
Why this matters more than a case fileIt is the one account where we own every number end to end, including the bad weeks. If the system did not work we would be the first to know and the first to stop paying for it.
Insurance

Fourteen live accounts, 90 days

Every insurance account we ran in the 90 days to 5 October 2026, anonymised, sorted by spend. $43,901 of ad spend, 797 enquiries, $55 blended. The last four rows are the honest part: accounts where the cost per enquiry is bad and we are either fixing them or telling the client to stop.

AccountAd spendEnquiriesCost per enquiry
Account A$5,982101$59
Account B$5,618153$37
Account C$4,390110$40
Account D$4,08871$58
Account E$3,920129$30
Account F$3,67553$69
Account G$3,26359$55
Account H$2,4737$353
Account I$2,35424$98
Account J$2,2714$568
Account K$1,7283$576
Account L$1,62339$42
Account M$1,28522$58
Account N$1,23122$56
Insurance, not mortgageThese are insurance agencies. Their cost per enquiry does not transfer to a mortgage audience, and we are not presenting them as if it does. They are here to show the operating range and that we do not hide the weak accounts.
What the weak accounts tell you

Three ways an account goes wrong

The offer is too narrow

Tight screens on a small market produce a handful of expensive enquiries. The fix is a wider screen or a bigger geography, not more budget.

Nobody works the enquiries

The most common failure, and not a media buying failure. If the first call goes out the next day, cost per enquiry stops being the problem.

How to check any of this

Open the accounts on a call

We will screen share the live ad accounts behind every row on this page, the funnel tracking behind the enquiry counts, and the CRM behind the funded files. Figures on this page were read from the accounts on 6 October 2026.