Nearly 4 in 10 property tax appeals that should have won are getting rejected — and it’s not because the case was weak. It’s because the evidence was generic. We pulled data from 10,000 appeals across our platform and worked with our engineering and case-management teams to find the pattern. Two issues accounted for almost all of it: appeals built on cookie-cutter evidence, and appeals that ignored the specific standards of the county appraisal district (CAD) reviewing them.
Here’s what we found, and what firms can do about it.
The Short Answer
Every time a property tax appeal is rejected, a firm loses billable outcome, not just one case — it’s the client relationship, the referral, and the time the agent already sunk into it. Our analysis of 10,000 appeals showed that when a case had a property tax appeal rejected, it was almost never because the underlying facts didn’t support a reduction. It was because the evidence package failed to meet the reviewing CAD’s actual expectations.
That breaks down into two root causes:
- The evidence was too generic — the same comp-selection logic and argument structure applied regardless of property or jurisdiction.
- The case ignored CAD-specific requirements — each appraisal district has its own tolerances, preferred comp criteria, and history of what reductions it will actually grant.
Why Property Tax Appeal Rejected Cases Follow a Pattern
We didn’t expect the data to be this clean. According to our CTO, the pattern showed up almost immediately once the team started segmenting rejected cases by cause rather than by outcome alone.
“We went in assuming rejections would be scattered — bad comps here, missing documentation there. What we actually found is that most rejections cluster around the same failure: the evidence wasn’t built for the specific CAD reviewing it. A generic strong case and a CAD-specific strong case can use the exact same property data and get completely different results.” — Sargis Ghazaryan, Chief Technology Officer
This matters because most appeal software — and most manual processes — treat evidence generation as a single, standardized step. Pull comps, build a packet, submit. But CADs don’t evaluate appeals against a universal standard. They evaluate against their own precedent.
Reason 1: The Evidence Is Generic, Not Personalized
A generic appeal treats every property the same way: pull nearby comps, apply a standard adjustment model, submit. It’s fast, but it’s also the reason a property tax appeal rejected rate stays high — reviewers can tell when a case wasn’t built around the property’s actual context.
Our project management team saw this play out directly in casework before it became a data question.
“Agents would come to us with cases they were confident in, and they’d still get rejected. Once we looked closer, the comps were technically accurate but not the comps that specific reviewer wanted to see. It wasn’t a data problem. It was a personalization problem.” — Vlad Evoyan, Head Project Manager
Reason 2: CAD-Specific Requirements Get Ignored
Every county appraisal district has its own point of judgment — what it considers a defensible comp, how much variance it will tolerate, and its history of what reductions it has actually approved. A case built without that context is, in effect, arguing to the wrong audience.
This is the piece most appeal workflows skip entirely, because it requires knowing each CAD’s behavior, not just its stated rules.
How We Fixed It: Location- and CAD-Aware Evidence
Once the pattern was confirmed, the fix followed directly from the two causes. Our algorithm builds each case around:
- The property’s exact location — not just the county, but the specific submarket and its comparable inventory.
- That location’s CAD credentials and history — the standards, precedent, and typical reduction range that specific appraisal district has actually granted in similar cases.
Instead of one evidence template applied everywhere, the system generates a packet calibrated to what the reviewing CAD is actually likely to accept — because it’s been built on that CAD’s own track record.
This is the same mechanism behind our Texas case study, where firms using CAD-aware evidence generation saw a 20x increase in income — not because they took on more clients, but because more of their existing caseload actually won, and the firm could process far more cases in the same time.
What This Means If You’re Evaluating Appeal Software
If your firm is losing cases you believe should win, the evidence itself is the first place to look — specifically, whether it’s built for the reviewing CAD or just built to look complete. A few questions worth asking of any tool or process:
- Does it account for which CAD is reviewing the case, or does it use one method everywhere?
- Does it use that CAD’s actual approval history, or general market data?
- Can it scale that personalization across hundreds of cases, or does it only work at a boutique, manual pace?
That last point is the one most firms underestimate. Personalized, CAD-specific evidence is straightforward to produce for ten cases by hand. It’s a different problem entirely at the volume growing firms need to hit.