Quickstart

View as Markdown Open in Claude Open in ChatGPT

A first run of Vintage Modeling takes you from creating your account, through uploading your first files, to reading your first curves. Each step below says what you will see and links to the page that covers it in depth, so you can follow the path now and come back for the detail later.

Before you start, have at least one export from your core system ready, ideally a set of monthly loan snapshots. Preparing your data explains what to export.

Set up your account and organization

Create your account. On the sign-up page, enter your name, your email, and a password of at least 8 characters, then press Create account. Any email address works. If you were invited by a colleague, the invite email links you to a sign-up page with your name and email already filled in.
Verify your email. Vintage sends a 6-digit code to the address you signed up with. Type it in and press Verify. The code expires after a short window. If it does not arrive, Resend code becomes available again 30 seconds after each send.
Create your organization. An organization is the workspace that owns your portfolio, your uploads, and your team. Give it a name (you can rename it later in Settings) and press Create organization. You become its first admin.

If you were invited, you skip this step: once your email is verified you are added to the organization that invited you, at the role on your invite.
Arrive in the product. A brief You’re all set confirmation appears once. An admin or editor whose organization has no loans yet lands on the Uploads page. Everyone else lands on Modeling.

Every onboarding step shows the account you are signed in as, with a Use a different account control if you need to start again under another email. Accounts and organizations covers invites, roles, and belonging to more than one organization.

Upload your first files

An upload moves through five stages, shown on a progress rail beside the screen: Choose Files → Remove PII → Upload → Column Mapping → Review.

Start an upload. On the Uploads page, press New upload.
Choose Files. Three cards name the file groups: Loan Snapshots, Origination Files, and Transaction Files. Click a card and add one or more files of that type. Choosing the card is how you declare the file’s type. Every file under one card must have the same column names (in any order). Vintage accepts CSV (.csv) and Excel (.xlsx) files. For a large portfolio, CSV is faster and lighter on your browser. Press Continue when your files are staged. See Choosing files.
Remove PII. Nothing has left your browser yet. For each file group, pick the Loan ID Column (required) and the Tax ID Column (optional; choose None if your file has none). Both are scrambled into one-way protected IDs before upload. Then tick any other columns that hold personal information, such as names, addresses, or phone numbers. Vintage pre-marks the columns that look personal, and the choice is yours to change. Press Confirm and continue. See Removing personal information.
Upload. Your browser sends the scrubbed files. Stay on the page until every file reads Uploaded, because the files can only be sent from an open tab. Once every file has arrived, the screen changes to say your files are being processed. You can close the tab at this point: Vintage emails you when the files are ready for column mapping, and the upload stays in Upload History so you can resume it. See Uploading and processing.
Column Mapping. Each column in your file appears in file order, with Vintage’s suggestion pre-filled. Map each one to a Standard field (a field Vintage understands, such as Current Balance or Charge-Off Amount) or keep it as a Custom attribute. Nothing is dropped either way. Some columns ask a follow-up question:
- Status, flag, and transaction-type columns list their values, and you say what each one means (for example, which Loan Status values mean Charged off).
- Date columns show the date format Vintage read; ambiguous dates default to US month/day order, which you can change.
- Percent columns such as Interest Rate ask whether 8.25 means 8.25% (Percent points) or whether 0.0825 does (Decimal fraction). Interest Rate defaults to Percent points.
- Amount columns that might be running totals ask whether they report each month’s activity, a life-to-date total, or a year-to-date total.
Loan ID is the only required field. The columns you chose as Loan ID and Tax ID at Remove PII are not listed again; a note names them, and they are saved with your mapping. Every other field is optional, and Review tells you what a missing one costs. Press Continue to confirm each file group. See Mapping columns.

Your first portfolio is finished by the Vintage team

Vintage sets up every new organization’s first portfolio itself. This is a service, not a check on you or your data. Until that setup is done, pressing Continue on Column Mapping saves your mapping and hands the upload to the Vintage team instead of opening Review. The screen says so, and tells you there is nothing to come back and finish.

While your first portfolio is being set up:

Upload History shows the upload as In verification, with a note that nothing there needs you.
Modeling, Portfolio, and Fields are not in the navigation yet. Uploads and Settings work as usual. You can keep uploading files, and each new upload is handed over at the same point.
A teammate from Vintage appears on your Team list as Vintage support. An admin can remove that member at any time.
If something about the upload needs your attention, the team hands it back with a note shown above your column mapping.

When the setup is done, Modeling, Portfolio, and Fields open, and every member of your organization gets an email saying the portfolio is ready, with a link to Modeling. See Your first upload.

Uploads once your portfolio is set up

From then on, uploads go through the same stages and end on the Review screen, which is the last check before you commit:

A verdict: Ready; Ready with gaps (the upload can be finalized, but some modeled output has under 50% of its loans ready); or Blocked (the upload contributes no usable loan data, so there is nothing to finalize).
A reconciliation of how your rows became loans, new versus already in your portfolio.
What you can model: how many loans are ready for credit loss, payoff, and prepayment, by which method, and what would bring more of them in.

Press Finalize upload to accept it. Vintage then folds the upload into your portfolio, and the confirmation says which of two things happened: the portfolio is already updated, or loans are still being folded in, in which case you can leave and Vintage emails you when the portfolio is updated. See Reviewing before you finalize.

Month after month, Vintage recognizes a file in the same format as last time and pre-fills your previous decisions, so a routine monthly upload is mostly a matter of confirming. See Uploading month after month.

Explore your portfolio

Open Portfolio. The summary band at the top counts your loans, snapshot observations, and transactions, and shows how many loans are ready for each modeled output. Below it, the loan table lists every loan by its protected ID, with its original amount, origination date, latest balance, and Status: Vintage’s own reading of how the loan ended (Paid off, Charged off, Exited (other)), or Open when your data shows no ending. See The Portfolio screen.

Define your first segment

The Segment Builder panel beside the loan table is where you build a segment: the slice of your portfolio you want to study. Each field you filter narrows the slice, and the live count beneath the panel title shows how many loans match.

A good first segment is one product type, then a narrower band within it. For example:

Filter a field that names the product, such as a custom Loan Type column, to one value: new auto loans.
Narrow by credit, with a FICO Score range of 660 to 719.
Narrow by vintage (the period the loans were made), with an Origination Date range of 2021 to 2023.

Filters on different fields must all match. Several values or ranges on the same field widen the slice. Vintage remembers your segment, and the same segment drives the Modeling screen. See Segments.

Read your first curves

Open Modeling. With your segment defined, the screen shows:

Credit loss: the cumulative loss curve by loan age, solid where your data backs it and dashed where it is projected, with the expected lifetime loss as a percent of original balance.
Prepayment: the speed curve, switchable between the monthly speed (SMM, single monthly mortality) and its annualized form (CPR, conditional prepayment rate), with bars showing how many loans could be measured at each age.
Pricing: the loan amount and term start from the segment’s own averages; enter your funding curve (the funds-transfer pricing, or FTP, curve: your cost of funds by term) to see the required rate and its build-up.

If the segment holds more loans than your organization’s modeling maximum, the screen says Segment is too large, names your maximum, and asks you to narrow the segment. Every chart and table has a Download menu for CSV or Excel. See The Modeling screen.