Plaid: Introducing a new categorization taxonomy and category list for personal finance management apps and tools


September 28, 2023

More powerful Transactions data for more personalized user experiences

Leah Karlins & Victoria Margolin

Today, we’re excited to announce several updates to our Transactions product, which enables financial services providers to do even more with transactional data. These new features and quality enhancements are designed to amplify your ability to provide tailored solutions and services, unlocking improved user outcomes. In one of our biggest updates to-date, Transactions now offers valuable new data fields (like merchant website and logo), a streamlined implementation of the Personal Finance Category (PFC) taxonomy, and improved data quality. Updates are available starting today for all Transactions customers using our latest API version 2020-09-14 (and the most recent version of our client libraries, if applicable).

Every day, consumers choose between thousands of digital apps and services in search of tailored financial solutions. In fact, a 2023 J.D. Power study found consumers report higher levels of satisfaction and engagement when they receive personalized financial advice. That’s why financial services providers, from fintechs to traditional financial institutions, are focused on delivering customized financial management tools that help consumers gain actionable insight into their spending patterns, receive personalized budgeting advice, and reach their specific financial goals. 

Transactional data is critical to delivering powerful user experiences. It offers deep insight into where and how consumers are spending their money, forming the foundation of not only budgeting tools, but also new consumer products and services like subscription management, low-balance alerts, and timely offers. In addition, transactional data can help protect consumers from short-term financial challenges, support consumers’ long-term financial health, and enable businesses to identify partnership opportunities. Industry-leading fintech providers and financial institutions already use Transactions including SoFi, Flexport, Betterment, and Varo.

→To learn more about our new Transactions updates, check out the API docs

More enrichments that paint a clearer picture of how and where consumers are spending 

Transactions already returns a breadth of transaction details including a cleansed merchant name, category, location information, and payment channel. Starting today, we’ll return several additional data fields, listed below, to help paint a clearer picture of how and where consumers are spending. Customers need to be on the latest API version 2020-09-14 to access the new fields, and customers using our client libraries should update to the latest version.

  • Counterparties: Insight into all the parties involved in a transaction, besides the merchant, including marketplaces (e.g. DoorDash), payment terminals (e.g. Square), or financial institutions (e.g. Amex). This data can be used to improve rewards payouts, identify partnership opportunities, and provide more granular spending insights.

  • Merchant Logo: The logo associated with the merchant, to aid user transaction recognition and build a more delightful user interface.

  • PFC Icon: An icon associated with the primary PFC which can be used when a merchant logo is not available, or consistently in place of a merchant logo if preferred. 

  • Merchant Website: The website associated with the merchant, also useful for user transaction recognition.

  • Entity IDs: A unique ID associated with common merchants and counterparties, making it easier to build custom rules or logic for handling specific merchants. Sometimes called a ‘merchant ID’.

  • Confidence Levels: A description of how confident we are that the provided counterparty is involved in the transaction (e.g. very high, high). We return this for counterparties and PFCs to provide more visibility into prediction confidence.

1Transactions API Response - Sample JSON
3"transactions": [
4 {
5 "account_id": "ybV8yjea61u8XjY5EXPLuEYRnV5MAwcOZbzOZ",
6 "amount": 28.34,
7 ...
8 "counterparties": [
9 {
10 "name": "Burger King",
11 "type": "merchant",
12 "logo_url": "",
13 "website": "",
14 "entity_id": "mVrw538wamwdm22mK8jqpp7qd5br0eeV9o4a1",
15 "confidence_level": "VERY_HIGH"
16 },
17 {
18 "name": "DoorDash",
19 "type": "marketplace",
20 "logo_url": "",
21 "website": "",
22 "entity_id": "YNRJg5o2djJLv52nBA1Yn1KpL858egYVo4dpm",
23 "confidence_level": "VERY_HIGH"
24 }
25 ],
26 ...
27 "name": "Dd Doordash Burgerkin",
28 "merchant_name": "Burger King",
29 "merchant_entity_id": "mVrw538wamwdm22mK8jqpp7qd5br0eeV9o4a1",
30 "logo_url": "",
31 "pending": false,
32 "pending_transaction_id": "NpZnO8VyJmu97X1qY7D8Hp6mbARE56fRaOV71",
33 "personal_finance_category": {
34 "primary": "FOOD_AND_DRINK",
35 "detailed": "FOOD_AND_DRINK_FAST_FOOD",
36 "confidence_level": "VERY_HIGH"
37 },
38 "personal_finance_category_icon_url": "",
39 ...
40 "website": ""
41 },
42 ...

Streamlined access to our latest and greatest categorization taxonomy

We’ve made it easier to get started with Personal Finance Categories (PFCs), our revamped categorization taxonomy rolled out last year. PFCs are optimized for a personal finance use case, with fewer categories than our original taxonomy and enhanced insight into the intent of a transaction. Plaid’s roughly 16 primary and 104 detailed categories are outlined in our API docs

Going forward, PFCs will be returned by default in the Transactions API response for all customers. While we’ll no longer actively support the category field which is based on the legacy categorization taxonomy, we’ll continue to return this field in the API response to ensure continuity for customers who’ve used it to build business-critical logic. However, we strongly recommend using PFCs to get the best categorization results. We are making ongoing investments in the PFC taxonomy, building on the greater than 90% accuracy available today.

Continued investment in quality improvements

We’ve been working behind the scenes to continue increasing fill rates and accuracy, listening to customer feedback, and analyzing the vast array of data signals offered by the Plaid Network. That includes the data fields that matter most to our customers: merchant name, category, and location. 

Alongside ongoing model adjustments, we’ve also made notable recent updates to our location data inputs. You’ll now see greater accuracy and fill rate for the existing lat, lon, address, and postal code fields, providing improved location details including geolocation coordinates that can be used to build new features like map widgets.  

Create more value by pairing Transactions with Enrich 

We know aggregated data is not the only source of transactional data. Plaid Enrich cleanses, categorizes, and enhances your internal transactional data from non-Plaid sources like branded cards or other banking products. For example, Credit Union of Texas uses Enrich to provide more personalized offers to their members and optimize the user experience - ultimately empowering their members to achieve their financial goals.

When paired together, Transactions and Enrich consistently enhance data across all sources of transactional data, delivering data parity and reducing time and resources spent on reconciliation. For end users, this consistency unlocks more seamless and intuitive user experiences. And particularly for financial institutions using Enrich to clean up their in-house transactional data, this clarity can also reduce disputes and their associated costs.

What’s next?

Plaid continues to invest in the quality of our Transactions product to help you build ever more innovative services and products for your end users. Interested in getting started with Transactions? Sign up on our Developer Dashboard or contact our team.

Already a customer? We’d love to hear how you’re using Transactions and welcome your feedback. Send us a note: