How Budgeting Apps Make Money

The aphorism, "if the service is free, you are the product," is just as true today as it was back in the 1970s when the idea was first expressed about TV (before cable and streaming, TV channels were broadcast for "free" to all viewers.) The advertising-based business model of early television and radio has never gone away. Rather, it has become more aggressive, more specific, and far, far more invasive.

There are many "free" options for personal finance management applications. When you use these "free" apps, the currency you are paying with is yur data: access to your spending patterns, your cash-flow, and information about you are paying down debt, saving for a house, or living paycheck to paycheck.

When Mint shut down in 2024 and Intuit pointed users toward Credit Karma, years of end user categorized spending did not just vanish into the ether. Rather, the data moved inside Intuit's product family, alongside credit reports, tax filings, and business accounting tools.

Free applications make money in the following ways (some of this is specific to personal finance apps, but much of it is general for all free apps):

• referral revenue from financial products,
• advertising and audience targeting,
• data licensing and analytics, and,
• subscriptions with export friction.

The first three depend on data leaving your device and entering an ecosystem you rarely see.

If you're not paying for it, you're not the customer; you're the product

The most user-visible way companies like Credit Karma make money is through referrals. By convincing you to click on a link for a financial product (a new credit card, a new loan), Credit Karma receives a cut of that sale. That third-party company used your data to target a specific sales offer at you, knowing you needed to consolidate debt, or refinance a loan.

Credit Karma's privacy policy states it does not sell personal information to unaffiliated third parties for their marketing lists. In this case, Credit Karma themselves are the ones combing through your financial data and presenting you with advertisements based on your current financial status. It does not violate their privacy policy because they are not a third-party. They can share your data with any product Intuit owns, and through any tracking details included in their Google analytics (who then, also, resell your data.)

Mobile finance apps add another layer. A CounterSpy benchmark of 160 App Store finance apps found that 64% declare cross-app tracking, 60% embed advertising SDKs, and 92% collect identifiers linked to your identity. Apple's privacy label definition of "tracking" includes sharing data with data brokers. Those numbers describe the category, not any single app; they still show how normal embedded analytics and ad code have become in finance software.

The middleman you didn't choose

When you tap "connect your bank," you usually authorize a financial data aggregator, not a direct pipe to your institution. Plaid and Envestnet/Yodlee sit in the middle. Plaid powers thousands of apps; settlement materials in In re Plaid Inc. Privacy Litigation cite roughly 5,000 mobile and web applications. Yodlee supplies personal-finance tooling to many of the largest U.S. banks and fintechs. The Wall Street Journal reported in 2015 that Yodlee also sells aggregated transaction data to investors and research firms.

Big Picture Money makes using these bank sync services optional for this reason. Whether you choose to use SimpleFIN, Plaid, or any other integration we add in the future, the vendor you choose to add is in your control. If you choose to use none of these services, Big Picture Money has robust import features that let you manually download exports from your bank and import them into BPM.

These bank sync services open many avenues where your data can be stored, potentially read by bad actors, and/or sold by these third-party companies to make extra money. It's not just the third-party companies themselves, either, it is also the third-parties that THEY use -- so whatever backend services Plaid or Yodlee are using.

Your data winds up becoming available to a wide-range of sources with this setup:

• The application itself
• Analytics apps
• Advertising partners
• Data aggregators
• Bank sync providers
• Enrichment services they provide
• Their third-party services.

Each of these services is a place where data can be stored, analyzed, or passed onward under terms you never read at connection time.

What the Plaid litigation established

A federal class action covering approximately 98 million Plaid users produced a $58 million settlement in 2022. Plaintiffs alleged Plaid collected more financial data than requesting apps needed and captured bank credentials through a Plaid Link interface designed to resemble bank login screens. Plaid denied wrongdoing. The settlement required data minimization, deletion of certain historical data, and Plaid Portal so users can view and revoke connections.

Plaid's current policy states it does not sell or rent financial data for advertising or marketing. That matters for accuracy. It does not mean your data stops at Plaid. The budgeting app on the other end of the connection has its own privacy policy, analytics stack, and business model. Plaid also permits use of aggregated and de-identified data to develop products, facilitate research, and assess service performance.

Yodlee's investor-facing business

The Wall Street Journal described Yodlee's side business in plain terms: data gathered from millions of daily card transactions flows to research firms that mine it for stock-market clues. Yodlee told the paper it could report "down to the day how much the water bill was across 25,000 citizens of San Francisco" or daily spending at a national restaurant chain.

In 2020, Senators Ron Wyden, Sherrod Brown, and Representative Anna Eshoo asked the FTC to investigate whether Envestnet/Yodlee sold consumer financial data without adequate consent. Envestnet disclosed a civil investigative demand from the FTC in SEC filings. Motherboard later published a leaked Yodlee client document describing how transaction rows are cleaned and delivered to buyers. Envestnet maintains the data is anonymized. Researchers and lawmakers have repeatedly challenged that claim.

Using an aggregator without becoming one

We are building Big Picture Money on a license-funded, local-first model. We are not in the business of selling your transactions or running referral ads against your categories.

We will offer Plaid through our optional Connected Services (coming soon).

But while bank sync through a third-party is a tremendous convenience, it is also a risk. When you choose to use these services (from SimpleFIN to Plaid to Yodlee or any others), you should be aware of what might be happening with your data.

Personally, I choose to use SimpleFIN and I sync my bank transactions with that. But each person should evaluate their own level of risk.

Big Picture Money makes it easy to choose which services you want to use, and we default to all third-party syncing being off when you first install the app.

How we are building the Plaid integration

We are very privacy and security conscious at Big Picture Money. This is why Connected Services are an optional add-on; you can choose not to trust us with your data, either. The umbrella of Connected Services, in general, refers to all the services where we need to store your data for any reason, even temporarily.

Plaid integration falls into that umbrella. Plaid, necessarily, needs to run on our servers to periodically sync data and to use our own method of authenticating to the Plaid servers. This means we need to pull your data and store it temporarily in our database.

We do our best to secure our servers and make sure no one can gain access -- but all data stored on the internet is a risk.

To mitigate that, when we store your data, we encrypt it with a key that is known only to your installed version of Big Picture Money on your computer. Once encrypted, we could not decrypt and read your data if we wanted to. A bad actor who somehow gained access to our database would only see a lot of garbage bytes that could never be decrypted.

When you run Big Picture Money and it syncs with our servers, that encrypted data is pulled locally to your machine, decrypted on your machine, and then stored in the secure, encrypted, database on your machine where all your other financial data is stored.

(note: this feature is still under development and is subject to change from the above description. When we release the full feature, we will release a thorough threat model as well.)

"We don't sell your data" is splitting hairs

Privacy policies draw careful lines between selling, sharing, using for product improvement, providing to service providers, and aggregating or de-identifying. Understanding the legalese in those privacy policies is more than most people are willing to wade through.

The claims these companies make are nominally true: Credit Karma won't sell your information to unaffiliated third parties, Plaid won't sell your financial data for marketing. Both statements can be true while your spending history still fuels a large adjacent industry.

Regulators have noticed the gap. In December 2024 the Consumer Financial Protection Bureau proposed a rule that would treat data brokers selling income, debt-payment history, or credit information as consumer reporting agencies under the Fair Credit Reporting Act, with permissible-purpose limits and accuracy duties. The CFPB withdrew that proposal in 2025. Consumer Reports called the withdrawal a step backward for consumers whose Social Security numbers and financial profiles continue to circulate through broker markets.

The FTC's 2014 study of nine data brokers remains the baseline scale reference: one broker held more than 1.4 billion consumer transactions and 700 billion data elements; another added more than 3 billion new data points each month. Few consumers know those companies exist.

Some of the places your transaction data can wind up

Below is an incomplete list of some of the third-parties where your data might wind up.

Card-linked marketing and purchase graphs

Affinity Solutions markets what it calls the largest deterministic consumer purchase dataset in the U.S., built from bank-direct debit and credit card feeds. The company claims coverage of more than 100 million consumers and tens of billions of verified transactions, enriched with geolocation, demographics, and partner data, then matched to digital IDs for ad targeting and campaign measurement. In 2026 Comcast Advertising announced a partnership to integrate Affinity's transaction data into its audience engine.

Oracle's acquisition of Datalogix showed the pattern earlier: offline purchase records linked to online ad platforms including Facebook and Google. Oracle later exited its advertising product suite, but the precedent stands. Purchases are used for advertisement targeting.

Credit bureaus and alternative data

Credit files are no longer the only lending input. Experian's Cashflow Attributes product sheet describes more than 940 attributes derived from bank transaction data: income regularity, expense categories, debt signals, and wealth indicators, built from what Experian calls depersonalized U.S. consumer transaction data processing on the order of a billion monthly transactions.

Experian also sells identity graph services that merge offline identifiers with digital behavior for targeting and fraud prevention. TransUnion's Neustar unit plays a similar role in the identity-resolution market.

Consumer Reports' 2024 study of Facebook data archives found Experian and TransUnion's Neustar among the most frequent companies sending event data to Meta. LiveRamp appeared in 96% of participant archives. Finance apps are one feeder among thousands, but bank-linked feeds are among the most sensitive.

Data brokers and identity-resolution vendors

LiveRamp, Acxiom, Epsilon, and Oracle's historical BlueKai business buy, supplement, and resolve consumer identities across channels. Major advertising holding companies acquired brokers outright (Interpublic Group and Acxiom; Publicis and Epsilon). Consumer Reports summarized the role: data brokers "collect and sell data about consumers with whom they often have no direct relationship."

Policy analyses of Credit Karma note that Intuit may supplement user profiles with data broker and public-record sources. You signed up for a credit score. The profile behind the score can draw from markets you never entered.

The downstream buyers are not all advertisers. The Brennan Center documented brokers selling personal information to predatory lenders, stalkers, scammers, and political consultants. A Consumer Reports study of people-search removal services found even paid deletion tools removed only about 35% of listed personal information after four months.

Investors and hedge funds

Hedge funds buy these data too. They use card swipes to guess a retailer's sales before the company reports them: how busy the stores were last week, how much people spent at a restaurant chain. Envestnet says the data is anonymized. The Wall Street Journal and Motherboard still found merchant names, amounts, and dates left in after the cleanup.

Ad-tech pixels on financial websites

Budgeting apps often ship as web dashboards plus mobile clients. The same tracking ecosystem that hit tax and mortgage sites hits finance products.

The Markup's Pixel Hunt project found major tax preparation sites sending income, refund amounts, filing status, and dependent information to Meta through the Meta Pixel. A 2024 Markup investigation of mortgage sites found more than 200 lenders sharing estimated credit, veteran status, occupation, and property details during application flows.

A congressional report spurred by that reporting concluded Meta and tax prep companies inappropriately shared millions of taxpayers' financial data for years. Meta told investigators it used the data for advertising and to train AI algorithms. A Treasury Inspector General audit confirmed consent procedures failed to identify specific recipients. Lawmakers urged the Department of Justice to pursue criminal enforcement in 2024.

Personal Finance applications rarely face the same congressional spotlight. The underlying mechanism (analytics pixels and SDKs firing while you enter sensitive numbers) is commonly used across many types of sites.

Breach dumps and concentrated infrastructure

Hosted finance stacks concentrate custody. When one banking-as-a-service provider fails, many brands feel it.

Evolve Bank & Trust disclosed in 2024 that a ransomware group accessed names, Social Security numbers, bank account numbers, and ACH transaction records for more than 7.6 million individuals, including customers of fintech partners such as Affirm, Mercury, and Wise. Evolve listed Plaid among its partners; reporting did not establish that Plaid's own customer database was breached. One breach in this shared financial infrastructure can expose data across dozens of services that all feel like separate brands to the user.

Sensitive inference from metadata alone

You do not need a broker to sell your data to learn a great deal from it. Recurring pharmacy charges, political donations, legal payments, fertility clinics, unemployment deposits, and buy-now-pay-later repayments each carry inference weight. Experian's cashflow product literature describes more than 100 expense categories derived from transaction text.

The Markup documented location brokers selling data sourced from prayer apps and dating apps to military contractors. Financial metadata carries parallel sensitivity: tithing patterns, clinic copays, and defense attorney retainers are all visible in transaction descriptions when someone holds the feed.

Four purchases can identify you again

Anonymization promises often assume stripping names and account numbers is enough. Research says otherwise.

MIT researchers Yves-Alexandre de Montjoye and colleagues analyzed three months of credit card records for 1.1 million people, anonymized to shop name, date, and amount. They reported in Science that four spatiotemporal points were enough to uniquely re-identify 90% of individuals in the dataset. Adding price information made three points sufficient in many cases. MIT's plain-language summary noted that three receipts plus a public Instagram photo could expose someone's full purchase history among a million records.

Senator Wyden's 2020 letter to the FTC about Yodlee cited that study against claims that transaction rows sold to investors remain anonymous. Your budgeting app knows you bought diapers at Target on Tuesday and paid a therapist on Thursday. Those are two points. The "anonymous" export was never anonymous in the way most people understand the word.

How brokers complete the picture of you

Data brokers and credit bureaus run identity graphs: systems that stitch fragments into a persistent profile. Experian describes merging digital and offline identifiers with its own datasets and client first-party data. Acxiom's Real ID platform uses deterministic and probabilistic matching to unify customer records across channels.

Intuit offers a concrete corporate example. TurboTax holds income, employer, and dependent data. Mint and Credit Karma hold spending categories and cash-flow patterns. QuickBooks holds business revenue for side gigs and small companies. Mailchimp holds engagement behavior. Their privacy policy language calls that "personalization."

Add web browsing from pixels, location from mobile SDKs, and credit-header data (name, address, date of birth, Social Security number fragments), and the portrait fills in. Consumer Reports found the average study participant had data sent to Facebook by 2,230 distinct companies across a three-year archive, from a sample universe of more than 186,000 company names. Financial apps are one input. They are a high-value input.

Signal visible in a transaction ledger	Common inference
Recurring pharmacy and specialist copays	Health and treatment patterns
Donations to clinics, advocacy groups, or religious institutions	Beliefs and life circumstances
Divorce attorneys, couples counseling	Relationship status
Job-search tools, unemployment deposits	Employment instability
Buy-now-pay-later repayments	Debt stress

Industry marketing materials describe category systems like these at scale. They do not need to read your private diary if they have your financial data.

Where it shows up in everyday life

Documented cases translate the map into moments people recognize.

A credit card offer arrives while you are quietly house-hunting. Mortgage inquiry categories, credit pulls, and browsing pixels can converge before you tell anyone your plans. Markup's mortgage investigation showed estimated credit scores and property details leaving lender sites during quote flows.

A "pre-approved" loan pitch matches your cash-flow stress. Experian's cashflow attributes explicitly target lenders who want transaction-derived income and expense signals beyond traditional credit files.

A retailer's ad seems to know you switched brands. Card-linked measurement ties ad exposure to verified purchase data. Affinity's business model is built on this.

An earnings preview moves a retail stock. Aggregated spend trends from transaction aggregators have fed investor research for years, as the Yodlee reporting established.

A scam caller knows your bank and general account balance. Data brokers sell identifiers and financial tier signals; breaches like Evolve add account numbers and ACH details to criminal marketplaces. The CFPB's withdrawn 2024 proposal cited brokers selling sensitive financial data to scammers as a core motivation for tighter rules.

Your tax refund amount shapes ads on social platforms. Markup and congressional investigators documented approximate income and refund data reaching Meta from tax filing flows, then reused for advertising and model training.

A Plaid connection you forgot three years ago still syncs. Visit my.plaid.com and you may find old apps you stopped using but never revoked. The 2022 settlement made that audit path explicit; using it is still your job.

Three practical checks: review Plaid Portal, review connected apps inside your bank's security settings, and read the App Privacy labels on iOS for finance apps you keep installed.

Choose incentives, not just features

Before you connect another account, ask a simple question: who gets paid when you use this app?

Follow the chain. You open a budgeting product. It pulls transactions through an aggregator. The aggregator stores copies, builds models, and serves other clients. The app runs analytics SDKs, shows partner offers, and may share data with corporate siblings or ad networks. Parallel markets buy card-linked feeds, cashflow attributes, and identity keys from brokers who never sent you a welcome email.

A license-funded, local-first tool flips the vendor incentive: you pay for software with cash, not with your data. Third-party aggregators still sit in optional sync paths, which is why we spell out the Plaid tradeoff above and keep SimpleFIN and imports available. We built Big Picture Money on that stack because household money deserves custody lines you can see.

If you want to understand what that looks like in practice, read how security and encryption work in the product. If the incentive alignment matches how you want to run your finances, see pricing.