Add commercial CA comparison group (7,138 organizations)

Matched on employee band, company info requirement, collection period and website popularity range. No domain overlap with the adopter file.

README.md

@@ -18,7 +18,16 @@ task_categories:
 language:
   - en
 size_categories:
-  - 1K<n<10K
+  - 10K<n<100K
+configs:
+  - config_name: default
+    data_files:
+      - split: train
+        path: Lets-Encrypt.csv
+  - config_name: commercial-ca-comparison
+    data_files:
+      - split: train
+        path: Commercial-CA-Comparison.csv
 ---
 
 # Let's Encrypt Adoption Among 5,000 Large Companies and Organizations
@@ -50,7 +59,7 @@ Let's Encrypt is a free, automated, and open Certificate Authority that provides
 
 ## Dataset Fields
 
-The data is in `Lets-Encrypt.csv`: 5,050 rows, 15 columns. Coverage is the share of rows where the field is populated.
+The adopter data is in `Lets-Encrypt.csv`: 5,050 rows, 15 columns. Coverage is the share of rows where the field is populated.
 
 | Column | Type | Coverage | Description |
 | ------ | ---- | -------- | ----------- |
@@ -72,6 +81,31 @@ The data is in `Lets-Encrypt.csv`: 5,050 rows, 15 columns. Coverage is the share
 
 The language and social columns are populated for fewer than 1% of rows.
 
+## Comparison Group
+
+`Commercial-CA-Comparison.csv` holds 7,138 large organizations that were found using a commercial certificate authority and were not found using Let's Encrypt. It is there so the adopter file can be measured against something rather than read on its own.
+
+It was built with the same filters as the adopter file: 5,001 or more employees, company information required, collected August 2026. It is also restricted to the same website popularity range, so neither file is truncated against the other. The two files share no domains.
+
+Columns are the 15 above plus `Certificate Authority`, naming the authority or authorities detected. 833 organizations were detected with more than one, written pipe separated.
+
+| Certificate authority | Organizations | Share |
+| --------------------- | ------------- | ----- |
+| DigiCert | 3,484 | 48.8% |
+| GlobalSign | 2,691 | 37.7% |
+| Sectigo | 1,445 | 20.2% |
+| GoDaddy SSL | 385 | 5.4% |
+
+Load it with `load_dataset("stackscan/lets-encrypt", "commercial-ca-comparison")`. The default config remains the Let's Encrypt file.
+
+### What the comparison supports
+
+It supports asking how Let's Encrypt adopters differ from large organizations that buy commercial certificates, across industry, country, size and founding year.
+
+It does not give an adoption rate. That would need a denominator of all large organizations, which neither file provides. Any percentage computed by dividing one file by the other is meaningless.
+
+Organizations can use more than one certificate authority. Those found with both Let's Encrypt and a commercial authority were removed from the comparison file, so the two groups stay separate at the cost of not representing mixed estates.
+
 ## Methodology
 
 Websites were analyzed for technical fingerprints associated with Let's Encrypt. An organization is considered a Let's Encrypt adopter when the relevant certificate or technology footprint is detected on its associated website.
@@ -98,7 +132,7 @@ This dataset can be used for:
 * Website infrastructure can change over time.
 * Results represent observations made in August 2026.
 * The dataset should not be interpreted as an official statement from any company regarding its technology stack.
-* Every row is an adopter, so the file describes what Let's Encrypt adopters look like. Measuring what share of large organizations use Let's Encrypt would require a comparable sample of non-adopters, which is not included here.
+* Every row of `Lets-Encrypt.csv` is an adopter. `Commercial-CA-Comparison.csv` provides a matched group that does not use Let's Encrypt, which supports comparison between the two but still does not support an adoption rate.
 
 ## Citation