When Ancestry.com’s Pro Tools Fail: A Professional Genealogist’s Experience with Ancestry’s Tree Checker Part 2

AI Generated

Last week, I shared my experience with most of Ancestry.com's Pro Tools—an add-on offered for $10/month (I got in for $7 with a promotional email). You can read about those features here. Today, I’ll dive into the tool that motivated me to subscribe in the first place: Tree Checker.

If you’ve noticed the new Tree Checker score on your Ancestry tree, you might be curious. Mine showed a 9.1—“Excellent.” That aligns with my belief that my tree is about 90–95% accurate. But let’s be honest: no one has a 100% accurate tree. Without DNA confirmation for every line, there will always be an element of uncertainty. Still, I’m committed to removing the detritus that’s accumulated over years of brick wall chipping, FAN Club research, and lineage society applications.

Back in 1990, I started my tree with 50 people, using a TI-89 cartridge program. By 1995, I was entering data into FamilySearch’s .paf format. When Ancestry came along, I uploaded my work to what’s now the ubiquitous .gedcom. My skills, and my tree, have grown significantly since then. I’ve cleaned up my Swedish, Croatian, French, German, Swiss, and Dutch branches. What remains is my largest line: Great Britain. Before writing my next book, I knew it was time to clean that section.

I regularly back up my Ancestry tree to other programs (Family Tree Maker, Legacy, RootsMagic), all of which offer tree-checking tools. But because Ancestry is my primary research platform, I’ve been hesitant to clean externally and re-sync. So I was hopeful that Tree Checker would finally give me an effective cleanup solution within Ancestry.

Here’s what happened.

Tree Checker: Expectations vs. Reality

When I launched Tree Checker from the dashboard, I was greeted with a gut-punch: 14,000+ possible errors.

The majority were labeled “People with no sources.” I immediately knew what was going on. In Ancestry’s early days, there wasn’t a “web link” option. I got around that by uploading source PDFs to the Gallery or by placing citations in the timeline. Unfortunately, Tree Checker ignores those, unless it’s housed as an official Ancestry “source,” it’s invisible to the system.

But that still left other problem categories:

• Possible Duplicates
• People with Only Tree Sources
• Other Possible Errors

I’ll add here that one of the most helpful “error types” isn’t even under Tree Checker, it’s found under Pro Filters > Family Lines > People Without Relationships. These are individuals floating without connections, often leftovers from attempts to delete a line. I had about 2,000 of these and quickly removed them.

People with Only Tree Sources was next. These were added from others' trees, unsourced. That’s an easy fix, either delete them or attach a hint. Done.

Possible Duplicates looked daunting at nearly 2,000, but the number was misleading. Triplicates and higher were counted separately, and after filtering, I had fewer than 1,000 to review. Some were legitimate merges (e.g., marriage records auto-adding a new spouse). Others were not duplicates at all: families who reused names after a child died, or multiple “Johann” Harbaughs with different middle names. I worked through them in two days.

The Glitches Begin

After carefully resolving every duplicate, I noticed something troubling: they didn’t disappear from the error list. No matter what I tried, refreshing, logging out, rebooting, clearing cache, Tree Checker continued to show errors I had already corrected. I even tried deleting and re-adding a person. No dice.

Still hoping for results, I moved on to “Other Possible Errors” and found myself stunned.

Ancestry itself was causing many of the flagged issues. For instance, if a child was born in 1937 and enumerated in the 1940 census, Tree Checker would flag it as “Resident listed before birth date.” The kicker? That census record was automatically added by Ancestry in the timeline for 1935. To clear the error, I had to delete 1935's entry FOR EVERY ONE born between 1936-1940.

Swedish church records were another problem. Ancestry indexes these by range (e.g., 1723–1728). If a child was born in 1724, Tree Checker flagged the 1723 record as occurring before birth. Completely illogical and a huge waste of my time to clean up!

Some new error flags also made no sense:

• “Birth/Death dates span more than 10 years” with only one sourced date. What?
• "Significant age difference between spouses" um, 2 years!
• “Marriage occurred after spouse’s death” when no death date was even given. See the above screenshot proving the error was false.

And here’s the real kicker: even when I corrected the problems, they remained in the count. Over the next three days, my “error total” would inexplicably rise despite spending hours cleaning.

Note that it says there are 2 possible duplicates but none show.

People with only tree sources shows 1 but none are provided.

Under all possible errors the counter states 201 but there is only 1 error showing and it is not an error when you go to that page.

This reminded me of a long ago problem Ancestry had with what was called "Ghost Hints." You can read my how-to-fix blog about it here. I tried that again but it appears that Ancestry has tightened up security and my fix it no longer worked.

As a genealogist, I dug into the data. I exported the report, analyzed the stats, and discovered something stunning:

75% of the Tree Checker results were false positives. That’s not a helpful tool -that’s noise! Seeing it graphically made me realize I had been sold a product that doesn't work:

To top it off, this also distorts your overall Tree Checker “score.” I now wonder what my real rating would be if the tool actually worked. With the changes that took, my score reached a 9.4.

The Final Straw

I then turned to the “No Sources” filter and began manually fixing issues from A–Be, X, Y, and Z. That’s when I hit the wall. Even attaching suggested Ancestry hints they no longer removed individuals from the list. Not user error, this was a flat-out malfunction.

That’s when I noticed the word Beta scattered throughout Pro Tools. Beta testing, by definition, is the final phase before a product goes public. Users test real-world functionality and provide feedback. But here’s the problem:

Ancestry released an untested tool to the public and then charged for it.

If you charge admission before the dress rehearsal is done, that’s not Beta testing. That’s profiteering.

But that's not all! Ancestry then sent me an email with their data about the changes I made to my tree:

What does 300% more duplicates found even mean?! We know I had no duplicates and most of those that they believed were duplicates were not. 84% fewer issues discovered? Does that mean my tree still has 16% undiscovered issues? If so, how would I ever find them when Pro Tools can't identify them and the counter doesn't work?

Where Do We Go From Here?

I’m not just disappointed, I’m concerned. It’s clear Ancestry recognizes the problem of flawed user trees. But their solution shouldn’t be charging extra for a broken product.

And now, they’ve rolled out something even more baffling, a $5,000/year “cohort club” promising professional coaching, a few DNA kits, and discounted branded merchandise. (Want a denim jacket? You can buy one at a discount.) Click the link as I'm not making this up. There VIP service, you get a working phone number if you have a problem. Pardon me, but I always thought that's what a legitimate business offered TO ALL OF THEIR CUSTOMERS for free.

In a recent webinar, the presenter said she came up with the idea of genealogy coaching. That’s interesting, since I’ve offered coaching on my website for over a decade at a fraction of the price. I believe everyone should have access to their heritage, not just those who can afford a luxury tier. I was also appalled to hear that professional genealogists charge tens of thousands of dollars a year. No, just no!

I use Ancestry daily and plan to continue. But I’ve cancelled my Pro Tools subscription.

Next week, I’ll share how I cleaned up my tree without shelling out extra cash.