One of my favorite pastimes is looking for discrepancies in filings! Much can be learned by following the additions, deletions or substitutions in filings as filings are living documents that tell a story and I look forward to more of these posts!
In my view, your 2 examples are a bit different, and I would go so far as to add a third category, though they all fall under the same umbrella.
1) Companies have been know to remove or change the wording from filings that they now perceive to be problematic.
2) There is also the instance, which you pointed out and which happens more than people would like to believe, where a company will deliberately change numbers they have reported in the past.
3) There is also the case where a company stops providing certain metrics that were once supposedly critical in analyzing their business but which are starting to become problematic.
I love pointing these things out when I can find them, since like you, I tend to believe there's a reason companies do these things, and that they will eventually matter.
On that note, here are 3 recent discrepancies i've pointed out.
1) IAS - the March 2023 quarter was the last quarter they disclosed their "total customer" count. This came after they posted 2100 total customers in Q1, 2103 in Q4 2022, and 2152 in Q3 2022.
2) GSHD - they recently combined reporting their TX-based and non-TX based franchisee count where they used to report them separately. One reason perhaps? TX-based franchisee count has been fairly steady, while non-TX franchisees had been in a freefall, declining 15% or so in the prior 3 quarters.
3) NAAS - this one's almost a three-fer. Changed their accounting method in Q1 2023 to remove the problematic "incentives paid" metric. This resulted in a restatement with prior revenues almost doubling on the removal of what was a contra-revenue item.
If you read through Paycom’s latest quarterly report, you would have noticed they clearly disclosed they have changed the way they calculate their revenue retention and disclosed the “new” numbers (with the new calc) back-dated to ‘23, ‘22 and ‘21 -(90%, 91%, and 94% respectively) so while there is a “discrepancy” between the 10Ks, that’s due to the change in calculation for these numbers that the company has since disclosed/corrected, however you want to frame it. Now, one could try and take issue with said change and argue there is more to it, but I think their explanation in the quarterly report makes sense. Cheers.
Well, red-faced and embarrassed... you're right, i missed that. Which raises another issue wit discrepancies, which I often go out of my way to find - when something "missing" or changed has merely been moved or in this case - explained. That said, as you point out - the bigger issue is "Why"?! Thanks for pointing that out.
Wondering if you or anyone has tried ChatGPT for this. I have found it very useful for summarizing, and I suspect it could be useful for finding discrepancies as well. I use the free version so there are limits, but still useful.
Good question - "ChatGPT spit out discrepancies." Of course, you would have triple check it. I'm sure that exists... if not it should. Might as well remove all levels of edge that might still exist. ;-)
Great piece and good examples. I complained (and wrote) for years that footnote, non-GAAP, and KPI revisions are difficult to track because—in contrast to restatements of high-level GAAP numbers—no rules prescribe how these revisions should be disclosed. With that in mind, many revisions are driven by SEC comments. Reading the narrative of the comment letters provides context to SEC's objections that triggered the revisions.
One of my favorite pastimes is looking for discrepancies in filings! Much can be learned by following the additions, deletions or substitutions in filings as filings are living documents that tell a story and I look forward to more of these posts!
Nobody is as good at finding them as you are!
100%
In my view, your 2 examples are a bit different, and I would go so far as to add a third category, though they all fall under the same umbrella.
1) Companies have been know to remove or change the wording from filings that they now perceive to be problematic.
2) There is also the instance, which you pointed out and which happens more than people would like to believe, where a company will deliberately change numbers they have reported in the past.
3) There is also the case where a company stops providing certain metrics that were once supposedly critical in analyzing their business but which are starting to become problematic.
I love pointing these things out when I can find them, since like you, I tend to believe there's a reason companies do these things, and that they will eventually matter.
Thanks Martin. Right you are - discrepancies come in various shapes/forms. No. 3 is so classic!
On that note, here are 3 recent discrepancies i've pointed out.
1) IAS - the March 2023 quarter was the last quarter they disclosed their "total customer" count. This came after they posted 2100 total customers in Q1, 2103 in Q4 2022, and 2152 in Q3 2022.
2) GSHD - they recently combined reporting their TX-based and non-TX based franchisee count where they used to report them separately. One reason perhaps? TX-based franchisee count has been fairly steady, while non-TX franchisees had been in a freefall, declining 15% or so in the prior 3 quarters.
3) NAAS - this one's almost a three-fer. Changed their accounting method in Q1 2023 to remove the problematic "incentives paid" metric. This resulted in a restatement with prior revenues almost doubling on the removal of what was a contra-revenue item.
I like this type of alert :)
Now I just need people to send in their faves.
If you read through Paycom’s latest quarterly report, you would have noticed they clearly disclosed they have changed the way they calculate their revenue retention and disclosed the “new” numbers (with the new calc) back-dated to ‘23, ‘22 and ‘21 -(90%, 91%, and 94% respectively) so while there is a “discrepancy” between the 10Ks, that’s due to the change in calculation for these numbers that the company has since disclosed/corrected, however you want to frame it. Now, one could try and take issue with said change and argue there is more to it, but I think their explanation in the quarterly report makes sense. Cheers.
Well, red-faced and embarrassed... you're right, i missed that. Which raises another issue wit discrepancies, which I often go out of my way to find - when something "missing" or changed has merely been moved or in this case - explained. That said, as you point out - the bigger issue is "Why"?! Thanks for pointing that out.
Wondering if you or anyone has tried ChatGPT for this. I have found it very useful for summarizing, and I suspect it could be useful for finding discrepancies as well. I use the free version so there are limits, but still useful.
Good question - "ChatGPT spit out discrepancies." Of course, you would have triple check it. I'm sure that exists... if not it should. Might as well remove all levels of edge that might still exist. ;-)
Great piece and good examples. I complained (and wrote) for years that footnote, non-GAAP, and KPI revisions are difficult to track because—in contrast to restatements of high-level GAAP numbers—no rules prescribe how these revisions should be disclosed. With that in mind, many revisions are driven by SEC comments. Reading the narrative of the comment letters provides context to SEC's objections that triggered the revisions.
Just takes a certain person who will wade through it.
That's me! I read all SEC comments and flag unusual disclosures. Happy to discuss further.
Definitely. Email me herb.greenberg@substack and we can hop on a call.