Creating a Study: The 3-Step Configuration Wizard

A study is created in one modal: the three-step configuration wizard. The entry point is New Study from the dashboard or the study list; the wizard runs in a full-screen panel with the three steps in a sidebar, and a dirty-check stops you from closing it with unsaved fields. The steps are ordered the way the run consumes them — identity and data first, context second, parameters last — and everything you enter is stored on the study, not just used for the run.

Step 1 — Study Info & Upload: identity and the data file

Field	Required	What it is
Study Name	Yes	The identity of the study — what reviewers see in lists and queues
Entity Name	Yes	The tested party — the entity being benchmarked
Financial Year	No	The period the benchmarking data covers, in a year or year-range form (2025, FY 2024-25)
Description	No	A free-text note on the study
Data File	Yes for the run	The comparables spreadsheet, dropped into the upload zone
Sheet Name	Yes once a file is attached	Which sheet of the workbook holds the data; the list is read from the uploaded file, and the control stays disabled until a file is behind it

The file uploads in this step, not at submission: you drop it, it uploads, and the file id is attached to the study when you submit. The formats, the upload flow, and the column-detection review are covered in Uploading Your Data File.

Step 2 — Study Context: the tested party profile

Field	Required	What it is
Tax Jurisdiction	Yes	The jurisdiction the study complies with (India by default)
Product / Service	Yes (a toggle)	Which side of the transaction is being tested; it also decides the function list shown below
Function Type	Yes, at least one	Multi-select of the tested party’s functions (e.g. Sales and Distribution, Marketing Support Service) — the input the FAR affinity scoring reads
Nature of Business	Yes, min. 30 characters	What the entity does, in prose — the text the qualitative screening matches comparables against
Screening scope	No	Optional structured scoping: Product Scope, Channel, Accepted and Excluded Business Models, Geographies, IP Ownership, RPT Profile, exclusion criteria and a materiality note

These fields are the tested party’s profile: the record the search and the screening read from. How the profile is used — and how the FAR characterization is derived from it rather than typed into it — is in Setting Up the Tested Party Profile.

Step 3 — Parameters & Run: the benchmarking parameters

Field	Default	What it is
Margin Selection (PLI)	OP/OC	The profit level indicator the study computes; the run is blocked if the dump lacks that PLI’s columns
Multi-Year PLI Averaging	Multi-Year (Golden Rule — Aggregate Raw Financials)	Whether the detected years are consolidated on the Golden Rule, or only the latest year is used
Arm’s Length Range (lower/upper percentile)	25 / 75	The percentiles that bound the range, with a one-click “apply jurisdiction convention” where the jurisdiction prescribes one
Tested Party Margin (%)	10	The tested party’s own PLI value — the position being benchmarked
Maximum RPT %	25	Rejects comparables whose related-party transactions exceed this share of revenue
Revenue Filters (min/max)	0 = unset	The size scope of the comparable population, with a currency and a scale unit (actual / thousand / million / crore) beside it
Employee Cost %	0 = unset	Rejects comparables whose employee cost falls below this share of revenue

Every parameter is recorded on the study — the record of what the run will do. The choices behind the PLI, the years, and the scope are in Choosing PLI, Years and Benchmarking Parameters.

What happens after you create it

Submitting the wizard posts the study with the file and the parameters:

1. The study row is written — identity, profile, parameters, file pointer, creator, tenant — and a creation entry lands in the audit log.
2. The analysis job is enqueued — the study walks Draft → Study Initialized → Processing, and the run starts in the background.
3. Progress is polled, not watched — the study page polls the job and shows the pipeline as eight steps: Queued, Preparing Data, Quantitative Analysis, Qualitative Screening, AI Screening (FAR Analysis), Web Research, Advanced AI Screening (Deep FAR), Final Analysis. Two of them are optional — Advanced AI Screening and Web Research — and both are marked (Not in your plan) when the tenant’s plan does not carry the single feature that gates them. When Web Research does not run, the fields it would have filled stay blank.
4. The study lands in review — when the run completes the state moves to In Review, the study appears in the manager’s queue, and the activity feed records the completion.

Each tenant has three analysis runs in flight at a time, against a global ceiling of ten; report jobs run on their own queue of five. If a queue is full, the submission is rejected with a message to retry in a few minutes — nothing is queued silently, and the study is not created.

Drafts, resumes, and re-runs

• Save as draft. You can keep a study as a draft without running it; the configuration is stored and the study sits in Draft until you are ready.
• Resume and run. A draft is re-opened in the wizard with its saved fields, the run is dispatched, and the same pipeline follows.
• Re-configure and re-screen. A run can be dispatched or re-dispatched while the study is in Draft, Study Initialized, Rejected or Failed: the parameters, the profile, the column mapping and the data file can all be changed, and the re-screen runs the pipeline again over the new configuration. The file itself can only be swapped in those same states — an in-review study’s results would desync from a new dump. From review onwards the configuration is frozen and a changed setup is a new study, not an edit.

FAQ

Which role creates studies? Any role can create one; the Analyst is the role that owns the pre-review work, and the study is assigned down the analyst → manager → partner chain as it moves.

Do I need the file before I can save? No — a study without a file saves as a draft. The file is required for the run.

What if the run fails? The study lands in Failed with the error recorded; re-submitting re-runs the pipeline. The state and the recovery paths are in Study Lifecycle.