Guide
Sift cleans a messy CSV or Excel file and gets it ready to import into a CRM, entirely in your browser. This guide walks through a full run: load, clean, reshape, score, map to your target, check it will import, and export. Nothing is uploaded, and you approve every change.
Sift runs 100% on your device, there is no account and nothing to install. Your rows never leave the browser, so it is safe to use with customer data you could not paste into an online tool or an AI chatbot. If you want to try it without your own data, use the built-in example dataset on the home screen.
Step 1
Drag a CSV, TSV, or Excel (.xlsx) file onto the drop zone, or click to browse. Sift parses and profiles it in your browser. For very large files (hundreds of thousands of rows), use Large file mode, which runs an in-browser database engine.
Step 2
The Audit tab shows a hygiene score (the share of cells that are present and have no detected issue), a per-column profile with the detected type (email, phone, country, postcode, date, number), and an outliers log listing values that do not fit their column. This is your before picture.
Step 3
The Clean tab proposes deterministic fixes, each with an exact count and a before-and-after diff you can review before applying: trim whitespace, standardize casing, normalize placeholder blanks, repair broken emails, normalize phone numbers and postcodes, standardize countries and dates, remove empty and duplicate rows. Approve the ones you want; nothing changes until you do.
Step 4
The Reshape tab restructures the data: combine columns (First + Last into Full Name), split a column by a delimiter, look up and join columns from a second file (e.g. add account IDs), filter rows, add a constant column to tag a whole list, or run fuzzy dedupe with survivorship to merge duplicate people into one golden record.
Step 5 (optional)
The Score tab lets you build a persona from weighted rules (role or seniority, industry, priority accounts, a numeric or recency window). Sift adds a Lead Score column and a transparent reasons column, then you extract the score range you want. Every number is explainable, no black box.
Step 6
In Match & check, drop your destination's import template (its column headers, ideally with a few sample rows). Sift auto-maps your columns to the target, and the readiness check flags blockers and warnings, required fields left empty, wrong types, values outside an allowed picklist, before you import, so you import once instead of three times.
Step 7
Download a clean CSV or Excel file. For CRMs that need accounts to exist before contacts (such as Microsoft Dynamics and Salesforce), generate a deduped accounts/companies file and import it first, then the contacts.
Step 8
Save everything you did as a named pipeline, then replay it on next month's list in one click. Pipelines match by column name, so the same clean-up runs again without redoing it.