I have been publishing a formation-to-dissolution ratio for El Paso County for a few months now. It is a simple number: count the businesses that formed in a year, count the ones that dissolved, divide. In 2024 it fell to 2.53 to 1, down from something closer to 4 to 1 in the good years, and that drop is the spine of most of what I have written about the local economy since.
Last week I went looking for the source of a spike. The monthly formation count for September 2025 came in at 2,021, against a baseline that had been sitting comfortably in the 1,300s and 1,400s. July 2026 did it again at 2,083. Two spikes, ten months apart, both roughly 45 percent above trend, both with no obvious story attached. So I went to find out what kind of businesses were forming.
They were forming at one address.
The spike that is not
Here is the monthly series with a second column added. The third column is the same count with a single principal address filtered out.
| month | all formations | at one address | ex-address |
|---|---|---|---|
| 2025-08 | 1,393 | 2 | 1,391 |
| 2025-09 | 2,021 | 667 | 1,354 |
| 2025-10 | 1,873 | 442 | 1,431 |
| 2025-11 | 1,214 | 2 | 1,212 |
| 2026-06 | 1,501 | 5 | 1,496 |
| 2026-07 | 2,083 | 598 | 1,485 |
| 2026-08 | 1,470 | 1 | 1,469 |
Read the last column top to bottom. It does not move. It is flat at roughly 1,350 to 1,800 with no spikes at all. Every apparent formation spike in the last two years is this one address, and when the address is quiet, the county's underlying formation rate is about as boring as a time series gets.
That is a genuinely unsettling thing to find in your own data. I had been treating those months as signal.
What is actually at that address
1155 Kelly Johnson Blvd, Suite 111, Colorado Springs. It is a commercial registered-agent suite, and it is the principal address on 4,495 El Paso County entities.
The agent organizations on file there are INCORP USA, INC. (1,771 entities), BIZGROW, INC. (912), MOUNTAIN BUSINESS, INC. (890), LEGACYVISTA CONSULTING INC. (464), and UNITED STATES CORPORATION AGENTS, INC. (21).
I want to be very clear about what this is, because the shape of the finding invites a story it does not support. These are ordinary, legal filing-service businesses. Registered-agent services exist because Colorado requires every entity to have an address in the state where legal service can be delivered, and thousands of people who form entities do not have one. Providing that address is the whole product. Nothing improper is happening at 1155 Kelly Johnson Blvd. I am not naming individuals, I am not speculating about who is behind any of these filings, and I have no evidence of misconduct because I went looking for a measurement artifact and that is exactly what I found.
The artifact is interesting on its own. These entities do not arrive evenly across the calendar. They arrive in two-to-four day clusters:
| window | formations at that address |
|---|---|
| 2025-09-23 to 09-25 | 667 |
| 2025-10-28 to 10-30 | roughly 442 in October total |
| 2026-07 (single month) | 598 |
Six hundred and sixty-seven entities in three days. That is not a wave of local entrepreneurship, that is a batch job.
The names read like a batch job too. The July 2026 cluster includes MERIDIANCOVE CONSULTING INC., BOLDMERIDIAN CONSULTING INC., COPPERVERTEX CONSULTING INC., AURORASPAN CONSULTING INC., OAKSUMMIT CONSULTING INC., TRUESUMMIT CONSULTING INC., QUANTUMHARBOR CONSULTING INC., BRIGHTCREST CONSULTING INC., and NIMBUSFORGE CONSULTING INC. The September 2025 cluster has the same programmatically generated shape, though with a different word list (LIMITLESS POWER, GRAZEFUL GAZE, VIBRANT PURPOSES, BRAIN WIZ).
One more structural tell. Entity type at that address skews hard toward corporations: 2,938 profit corporations against 1,485 LLCs, in a county whose overall mix runs roughly six LLCs to every corporation. That inversion is what drives the apparent surge in profit-corporation formations I would otherwise have written up as a shift in how local founders are organizing.
Why the bias only runs one way
If these filings were noise, they would cancel out. They do not, and the reason is in the status column.
| formed | Good Standing | Dissolved | Delinquent | total |
|---|---|---|---|---|
| 2022 | 1,222 | 6 | 8 | 1,236 |
| 2023 | 483 | 4 | 9 | 496 |
| 2024 | 613 | 13 | 9 | 636 |
| 2025 | 1,131 | 11 | 4 | 1,148 |
| 2026 | 614 | 1 | 0 | 615 |
Of the 1,148 entities formed at that address in 2025, 1,131 are still in Good Standing. Eleven dissolved.
Every one of those 1,148 entered the formation numerator the day it was filed. Almost none of them will ever reach the dissolution denominator, at least not on any timeline that matters to a ratio I publish quarterly. An entity that never trades and never files a dissolution just sits there in Good Standing, permanently counted as a birth and never counted as a death.
So the bias is not noise. It is one-directional and it pushes the ratio up.
What it does to the published series
Here is every year in the published series with an ex-address column beside it.
| year | formations | at that address | ex-address | dissolutions | published ratio | ex-address ratio | address share |
|---|---|---|---|---|---|---|---|
| 2019 | 11,499 | 22 | 11,477 | 2,938 | 3.91 | 3.91 | 0.2% |
| 2020 | 12,860 | 24 | 12,836 | 3,123 | 4.12 | 4.11 | 0.2% |
| 2021 | 15,142 | 31 | 15,111 | 3,521 | 4.30 | 4.29 | 0.2% |
| 2022 | 18,824 | 1,235 | 17,589 | 4,584 | 4.11 | 3.84 | 6.6% |
| 2023 | 19,347 | 496 | 18,851 | 5,007 | 3.86 | 3.76 | 2.6% |
| 2024 | 16,279 | 636 | 15,643 | 6,437 | 2.53 | 2.43 | 3.9% |
| 2025 | 18,615 | 1,148 | 17,467 | 5,487 | 3.39 | 3.18 | 6.2% |
| 2026 (through Sep 3) | 13,180 | 615 | 12,565 | 3,827 | 3.44 | 3.28 | 4.7% |
Look at 2019 through 2021. The address contributes 22, 24, and 31 formations, which is a rounding error, and the two ratio columns are identical to two decimal places. Then 2022 arrives with 1,235 and the columns separate for good.
That is the part that matters. This is not a constant offset I can wave off, it is a structural break sitting in the middle of a series I have been comparing across years. The pre-2022 numbers are clean. Everything after is carrying a variable amount of filing-service throughput, between 2.6 and 6.6 percent depending on the year.
I want to be precise about the 2024 figure I have published repeatedly. 2.53 to 1 is correct. It is the correct value for the stated definition, and that definition counts every qualifying filing in the county, which is the right definition for a statistic about business formation as the state records it. The 2.43 is a second read on the same year with agent-suite filings stripped out. This is a qualification, not a retraction. Both numbers are true, they just answer slightly different questions, and from here I intend to carry both.
Where that leaves the recovery
The story I have been telling is that 2024 was the trough at 2.53 and the county has been climbing back since: 3.39 in 2025, 3.44 so far in 2026. Removing the address does not erase that recovery, but it does flatten it. Ex-address, the trough is 2.43 and the climb runs to 3.18 and 3.28. Still up, just less dramatically, and part of the gap between those two paths is a filing service's order book rather than the county's founders.
There is a cleaner way to check. Comparing a partial 2026 against full prior years is bad practice anyway, so here is January through August only, the same eight months in each year, on the full published definition:
| Jan-Aug | formations | dissolutions | ratio |
|---|---|---|---|
| 2023 | 14,360 | 3,228 | 4.45 |
| 2024 | 11,388 | 3,806 | 2.99 |
| 2025 | 12,046 | 3,687 | 3.27 |
| 2026 | 13,062 | 3,769 | 3.47 |
Like for like, the recovery holds. 2.99 to 3.27 to 3.47 across three years, and the mechanism underneath it is the healthier of the two possible ones. Formations are up 8.4 percent year over year while dissolutions are essentially flat, up 2.2 percent. The ratio is improving because the numerator is growing, not because the denominator collapsed. That distinction survives the address problem, and it is the single most reassuring thing in this entire analysis.
What does not survive is any month-to-month reading. If someone points at September 2025 or July 2026 and calls it a surge in local business formation, that is 667 and 598 filings from one suite number talking.
The rest of the list
1155 Kelly Johnson Blvd is the big one, at 1,737 formations from January 2025 through August 2026, but it is not alone. Three other addresses cleared three figures over the same window:
- 5142 N Academy Blvd Unit #4322, with 224 formations
- 16055 Old Forest Pt Ste 301, with 144
- 6547 N Academy Blvd #2558, with 90
Individually none of those moves a county-level statistic. Together with the main address they are enough that I now check concentration before I report any month as unusual.
One practical note for anyone reproducing this. The suite line in the Colorado Secretary of State data is not normalized. Suite 111, SUITE111, SUITE 111, Ste 111, and NULL all appear on records at the same address. Match on the street portion, never on the suite field, or you will find a fraction of what is there and conclude the effect is small.
I will say it once more, because a finding shaped like this one gets misread easily: nothing improper is occurring. These are legal services doing legal work, and the entities they file are real filings that the state really recorded. The problem is entirely mine. I built a county-level indicator out of a filing count, and a filing count cannot tell the difference between four thousand people starting businesses in Colorado Springs and four thousand mailboxes sharing a suite number. That distinction has to be imposed from outside the data, which means it is a decision I make and disclose, not a fact the database hands me.
Every number here comes from the Colorado Secretary of State business entity and transaction files, current through September 2, 2026, filtered to the 21-city El Paso County list I have been using since the first piece in this series.