Where the money actually leaks
Most wasted budget in a property funnel is not lost in the auction. It is lost after the click: long forms, duplicated conversions, and audiences that keep paying to reach people who have already bought. Fix the measurement first, or every later decision is guesswork.
This is counterintuitive for teams who have spent years optimising bids, and it is why the optimisation plateaus. Auction efficiency has a floor set by your competitors. Post-click waste has no floor, because nobody is competing with you to fix it.
How to find your own leaks
Three numbers will tell you most of what you need, and all three are available this afternoon. Compare the conversion count in your ad platform against the count in your CRM. Compare form starts against form completions. And check what share of retargeting spend is reaching people who already converted.
Where the first two numbers disagree by more than a small margin, your measurement is wrong, and every optimisation decision built on it has been guesswork dressed as analysis. That is the first thing to fix, because the other two numbers cannot be trusted until it is.
The form is the biggest single lever
Eleven fields on a first-touch enquiry form is a filter, not a qualifier. Ask for name, phone and the one thing your sales team genuinely triages on. Everything else belongs in the follow-up call, where a person can ask it properly.
The objection is always that sales needs the information. They do, and they will get it on the call, from somebody who can ask a follow-up question and hear a hesitation. A form field can do neither. What a long form reliably does is remove the people who were curious but not yet committed, which in property is most of the pipeline.
There is a version of this worth defending. A deliberately long form, used as a qualifying filter, with the cost in volume accepted and measured, is a legitimate strategy. It is rarely what is happening. Usually nobody decided: the form grew a field at a time, because each team wanted one.
Tracking that fights itself
Multiple tracking scripts installed over years by different agencies will each claim the same conversion. The platforms then optimise towards inflated numbers, the reporting shows growth that did not happen, and budget follows the inflation.
The fix is tedious rather than difficult. Inventory every script on the site, decide which one is authoritative for each event, remove the rest, and deduplicate on a single lead or transaction identifier. Expect reported conversions to fall. That fall is the measurement becoming honest, and it needs explaining to whoever reads the weekly number before it happens rather than after.
Retargeting people who already bought
Converted-audience exclusion is a checkbox, and it sits unticked on a remarkable number of accounts. In a category with a long consideration window and a high value per sale, spend reaching buyers who have already signed is pure loss, and it stays invisible unless somebody goes looking for it.
The same applies to audiences built on all site visitors, which quietly include your own staff, your agency, and every job applicant who read the careers page. Exclusions are unglamorous and they recover budget immediately, which makes them the best-value hour in the account.
What we would not touch
Two things get reached for first and should be left alone until the above is done. Bid strategy is the obvious one: changing it while the conversion feed is duplicated just teaches the platform to optimise harder towards a number that is wrong. Fix the signal, let it stabilise for a fortnight, then revisit bidding with data worth optimising against.
The other is the site rebuild. A rebuild is the most expensive available response to a funnel problem and the slowest to produce an answer, and it usually arrives with new tracking that resets whatever history you had. If the form, the measurement and the exclusions are wrong, they will still be wrong on the new site, and you will have spent a quarter finding that out.
What to fix in the first fortnight
Deduplicate the tracking, shorten the form, exclude converted audiences from retargeting, and put a single honest number in front of the weekly meeting. None of it needs new budget, and all of it is measurable inside a month.
The order matters more than the list. Measurement first, because everything after it depends on numbers you trust. Then exclusions, because they recover spend immediately and cost nothing. Then the form, because it changes volume and you want clean measurement in place before you move it. Doing the form first is the common mistake, and it leaves you unable to prove whether it worked.