Skip to main content

For analytics products

Analyze what used to be plain text.

Analyzing transaction data takes stable dimensions. TXNORA derives merchant, category and labels from raw descriptions, so you can work with them in your own reporting.

Request trial accessSee before and after

The starting point

Raw text is hard to group.

The transaction description alone is not enough to report by merchant or category. References like the one in “AMZN MKTP DE*X4H88291” or additions like “REWE SAGT DANKE 1834” make lines unique, even when they may belong to the same merchant.

Mapping tables and text rules of your own only help as long as someone maintains them. The effort goes into the data foundation instead of the analysis itself.

Raw descriptions as they arrive

  • PAYPAL *ADOBE – PAYPAL: Payment provider; ADOBE: Merchant

  • REWE SAGT DANKE 1834 – REWE: Merchant; SAGT DANKE: Extra text; 1834: Number

  • AMZN MKTP DE*X4H88291 – AMZN MKTP DE: Abbreviations; X4H88291: Reference

Example data. The breakdown explains the raw data and is not API output.

What structured context changes

Structure you can compute with.

Three building blocks for analytics on transaction data.

  1. Group by merchant

    A normalized merchant name turns many raw lines into one dimension you can group, count and compare by.

  2. Category and labels as axes

    Category and labels give you two levels for segments and filters: one consistent main category plus additional context.

  3. Confidence as a quality signal

    The confidence value travels with every result, so you can decide which values go into a report and which to treat separately.

Before and after

Free text becomes dimensions.

Three typical transactions as a data foundation: first as free text, then with merchant, category, labels and confidence as fields you can analyze.

Data for your analysis

Three example transactions

  1. Raw

    PAYPAL *ADOBE

    Structured

    Merchant: AdobeCategory: Software

    Amount: −59.49 EUR

  2. Raw

    REWE SAGT DANKE 1834

    Structured

    Merchant: REWECategory: Groceries

    Amount: −42.17 EUR

  3. Raw

    AMZN MKTP DE*X4H88291

    Structured

    Merchant: AmazonCategory: Retail

    Amount: −129.90 EUR

Example data from three typical transactions.

The confidence values are examples, not measured accuracy.

Relevant fields

What your analysis needs.

For analytics, what matters are fields you can group and filter by, plus the confidence of each result.

  • Merchant

    merchant.name

    A dimension to group and compare by.

  • Category

    category.id · category.name

    A consistent main axis for segments.

  • Labels

    labels

    An extra level for filters and segments.

  • Confidence

    confidence

    The basis for treating uncertain values separately.

POST /v1/transactions/classifyExample response
{  "transaction_id": "txn_demo_001",  "merchant": {    "name": "Adobe"  },  "category": {    "id": "software",    "name": "Software"  },  "labels": [    "SOFTWARE",    "DIGITAL_SERVICE"  ],  "confidence": 0.984}

Relevant for this use case

Simplified example.

More use cases

Other products, the same foundation.

  • 01

    Fintechs

    Make transaction feeds easier for people to understand.

    PAYPAL *ADOBEbecomesMerchantAdobe

  • 02

    Accounting platforms

    Add merchant and category context before transactions enter a review workflow.

    REWE SAGT DANKE 1834becomesCategoryGroceries

  • 03

    Expense tools

    Bring more consistency to spending categories and review queues.

    PAYPAL *ADOBEbecomesLabelsSOFTWARE · DIGITAL_SERVICE

Test TXNORA with your product.

With free trial access, you see which fields matter for your use case. After your request, you receive access to the API, the developer documentation and our pricing.

Request trial accessAll use cases