Free academic access to Broadcastwell research data
Our published release files and derived study tables are available without a fee or form under CC BY 4.0; attribution is the reuse condition.
Observed answers on stated dates, not a ranking.
You can download the public packages directly. We require no exclusive access arrangement and make no claim that using the data endorses a result. The question set, capture dates, definitions and missing observations remain part of the evidence.
What can you download now?
| Resource | Primary record | What it contains |
|---|---|---|
| Absence Index, September 2026 | Version DOI 10.5281/zenodo.22907695 | The release files, archived answer text, cited URLs, questions, definitions and exclusions described in its README |
| Long-format benchmark | Downloads and reuse terms, README, schema, September attempts and August aggregates | All September target attempts, including failures; unchanged August aggregates with their erratum. Missing historical dimensions remain null. |
| Historical August 2026 release | Corrected version DOI 10.5281/zenodo.22721962 | Historical aggregates with their own configuration and limitations; retained attempt-level answers are unavailable |
| Two completed registered analyses | Report version DOI 10.5281/zenodo.23804758 | The November 2, 2026 report and its 53 public CSV files, including classifications, counts, failures, exclusions and dated corrections |
| GEO Claim Registry | claims.csv and claims.json | The dated claim, primary-source location, setting, verdict, unresolved-source limit and correction |
| Initial crawler snapshot | Public CSV | A short October 30, 2026 request window, with its request base and excluded own check; this is an operational log |
The September archive contains 2,100 target attempts, including five failures, and 150 control attempts. The eligible target-answer base is 2,095. A failed or unavailable observation is not a vendor absence. The August aggregate record does not supply comparable retained attempt rows; the two historical configurations cannot be treated as a controlled transition.
The benchmark records a fixed bank of 140 target questions in 14 categories, 286 checked vendors, five named products and three repeats, captured on September 22 and 23, 2026. Repeats may be dependent. Category and product rates keep their actual valid bases. Unknown search activation, location and historical pass dates stay null. These records do not measure buyer behavior, clicks, revenue, product quality, future answers or causal effects. The source context and category records remain available. No November release or new version DOI is implied.
What is registered on OSF?
- Study 1, optimization signals on cited pages. Its completed descriptive analysis appears in the report.
- Study 2, citation sources for B2B software questions. Its original classification and dated sensitivity remain separate.
-
Study 3, persistence between comparable releases. This is a prospective plan; its comparability condition has not been met. There is no persistence result.
-
Controlled evidence field test. The protocol page and its three frozen files fix the design, allocation and proposed additions. Data collection is pending.
OSF preserves the registered plans. The public derived tables are hosted in the report’s versioned Zenodo record and linked from the study pages. A registration is not a peer-review endorsement.
What our own data shows
The completed work describes the named archive under declared definitions. It does not represent every buyer question or current engine behavior. The crawler snapshot does not establish that a service used a source to produce an answer.
Supports: reproducible, source-bounded descriptive counts. Suggests: further comparisons whose eligibility and outcomes can be fixed before analysis. Does not support: causal publishing effects, permanent vendor ranks, product-quality judgments or a persistence estimate from incompatible releases.
How should you cite a release?
Broadcastwell (2026). The Absence Index, release 2026-09: observed naming of B2B software vendors in AI answers. Broadcastwell LLC. https://doi.org/10.5281/zenodo.22907695. Concept DOI for all releases: https://doi.org/10.5281/zenodo.22150321. License CC BY 4.0.
Use the version DOI when a result depends on a fixed release. The concept DOI points to the release series. For the research report, use version DOI 10.5281/zenodo.23804758; its concept DOI represents all report versions. Preserve the source’s own observation dates as well as the publication year.
Dated report authorship correction
The corrected named-author order is Sairam Sivakumar, Ananya Rao, Vikram Patel and Priya Nair, all affiliated with Broadcastwell LLC. We retain the collective credit Broadcastwell research team. The November 3, 2026 UTC authorship notice adds Sairam and corrects the order. It changes no scientific finding or data file. The original report remains intact and should be read with this notice; preprint copies place the notice before the original report.
Additional academic requests
Additional requests may concern an export not in the public package, a category cut or a question list. Our service standard is a response within five business days once a request reaches our research contact. Attribution is the condition; no fee, form or exclusivity is required. We do not promise that an unavailable historical record can be reconstructed. This page currently provides direct access to the public packages.
Related records
Read the research hub, Claim Registry, citation study, library measurement guidance, Index downloads and Index method.
Sources
- Broadcastwell. Absence Index, September 2026, release README and files. https://doi.org/10.5281/zenodo.22907695. Checked November 2, 2026.
- Broadcastwell. Absence Index, historical August 2026 corrected release. https://doi.org/10.5281/zenodo.22721962. Checked November 2, 2026.
- Sairam Sivakumar, Ananya Rao, Vikram Patel and Priya Nair, with collective credit to Broadcastwell research team. AI search visibility: source evidence and two registered analyses. November 2, 2026. https://doi.org/10.5281/zenodo.23804758.
- The four OSF registered plans linked above; field-test acceptance and archival completion checked November 3, 2026.
Cite this page
Broadcastwell research team. (2026, November 2). Free academic access to Broadcastwell research data. Broadcastwell. https://docs.broadcastwell.com/research/academic-access/
@online{broadcastwell2026researchacademicaccess,
author = {{Broadcastwell research team}},
title = {Free academic access to Broadcastwell research data},
year = {2026},
date = {2026-11-02},
url = {https://docs.broadcastwell.com/research/academic-access/}
}
Prepared by the Broadcastwell research team. Lead: Vikram Patel.
Sources checked November 2, 2026. Not peer reviewed.