# Review earnings: what changed?

Lululemon is the worked example. The task is to distinguish consumer weakness, product execution and competitive pressure using comparable financial evidence. It is useful for preparing an earnings review or deciding which questions need further investigation.

- [Open the full interactive report](report/)
- [Download the editable financial table](earnings-update.csv)
- [Open the illustrative thesis checkpoint](thesis-check.html)
- [Original report instruction](prompt.txt)
- [Adapted earnings-update prompt](earnings-prompt.txt)
- [Adapted thesis-check prompt](thesis-prompt.txt)
- [Full presentation data and evidence](report/report-data.json)

## Inputs and cutoff

Research cutoff: **September 4, 2026**. Lululemon FY2021–FY2025 and Q1/Q2/H1 FY2026 versus matching prior-year periods. Thirteen SEC, issuer and investor sources; Nike, On and Athleta comparisons retain their distinct reporting periods. Alo and Vuori appear only where public evidence supports the discussion.

The original task supplied a research question, not an existing spreadsheet or a prior investment thesis. No customer account, live feed or paid consensus terminal was used. New filings after the cutoff are not reflected here.

## What you can use

The report contains the deterioration timeline, regional comparisons, earnings and cash-flow bridges, peer disclosures and illustrative recovery scenarios. Underlined metrics expose sources and calculation inputs. The CSV exports existing reported and calculated facts with fiscal periods, units, IDs and source URLs. It is a data table, not an Excel model with formulas or an automatic model-update integration. Signed capex is a cash outflow; currencies remain in their recorded units.

The earnings prompt is adapted for reuse. It was not the exact prompt that produced this historical report. The thesis checkpoint is a **retrospective illustration** whose assumptions were written after reviewing the saved analysis. It is not a historical prediction, customer trial, tracked portfolio or investment-performance claim.

## Run this on your company

Open the FinRunbook repository in your agent. Copy an adapted prompt above, replace the company and cutoff, and supply any real KPI template or dated thesis you want retained. The agent runs the research and review workflow; viewing this static example does not run the skill. No savings, completion speed or repeat-use rate has been measured for this example.

## Evidence boundaries

The report's refund exclusion is a narrow analyst adjustment, not a fully normalized earnings series. Standalone markdown effects, exact customer switching and private-company financials are not invented. Actuals, guidance and scenario assumptions remain separate.

Original validation, model and browser receipts retain their September 4 scope and timestamps. They describe author-led research and checks, not an independent audit. The later publication review covers packaging, exports and the instructional page. Financial arrays and conclusions are unchanged. Raw issuer documents, source archives, machine paths and working scripts are excluded; source URLs and capture identifiers/hashes remain for traceability. Public factual extracts do not supply a third-party data license.

[Back to all examples](../index.html)