We started Heron Data to simplify the process of understanding an SMB's financial health. In early 2020 it was the midst of the Open Banking revolution - an age where financial data could seamlessly move from antiquated institutions to wherever the data subject wanted it to be. Plaid, Finicity, TrueLayer, and a roster of other players were founded to make that data transfer easy. We thought that the age of downloading and sending three months of bank statements as part of your loan application were over. How wrong we were!

When we started Heron we only accepted data in text format (JSON / CSV / etc.) but every day we heard from our customers that applicants banking with regional banks or credit unions had issues connecting via API aggregators because 'the institution is not supported'. Some applicants are simply unwilling to give access to their bank account to a tool they might not trust. Many traditional lenders, especially those who receive leads through brokers or ISOs, have a primarily PDF-based dealflow. Even for the most tech-forward lenders PDF statements make up a significant minority of applicant volume. Our customers were faced with the choice of ignoring PDF-based deals or manual processes which limited automation.

After six months in closed beta - we are excited to launch bank statement parsing and fraud detection! You can now send Heron PDF bank statements in addition to data pulled through API aggregators (you can even send data from multiple sources for the same company) and we will carry out fraud checks and parse the statement's transactions, balances, and other key data points.

We process ~90% of bank statements with 99%+ accuracy without a human-in-the-loop. The whole process takes less than a minute and we'll tell you up front if we can't parse the document successfully so you can make other arrangements. After parsing the data, we enrich it and you can view the results in our intuitive Dashboard or calculate Metrics about the company like true revenue, balance analytics, NSF days, DSCR, and much more. If you share your underwriting criteria (e.g. a minimum revenue requirement, minimum average balance, # of NSF days for example), we will tell you in real time whether the applicant qualifies so you can make sure your Underwriters only spend time on opportunities that are likely to convert.

If you'd like to join 100+ other B2B Lenders and Fintechs in automating bank-statement based workflows with Heron or are interested in testing out our solution, book a demo or drop us an email at sales@herondata.io.

Heron Data launches Bank Statement Parsing & Fraud Detection

Brendan Coons is the Director of Risk at Torpago, a disruptive financial technology solution empowering companies by modernising their credit and spend management. Torpago chose Heron Data to empower their Credit team to maximise their revenue and decrease losses by enriching banking data for their underwriting and portfolio management processes.

50% Increase in Revenue resulting from dynamic limit setting

Up to 96% Underwriting Time Saved from "received" to "approved"

Decreased Delinquency Rates Below 1% for 2022 approval vintage, by:
1. More accurately assessing applicants in the initial underwriting stage, by rejecting cases of suspected fraud and underwriting to a higher cash coverage ratio (increased by up to 10x)
2. Monitoring the portfolio to decrease credit limits and exit earlier, for customers whose financial position is degrading

Prior to using Heron, the Torpago team was restricted in three key ways:

Low Data Diversity: A low number of key financial data points they could utilise to make a decision for their applicants’ non-Torpago accounts

Manual Summaries: Everything needed to be manually pulled together into a spreadsheet to recreate a simple cash-based P&L, which required a high level of manual effort for non-ideal results

Hard to Monitor Over Time: Portfolio monitoring and dynamic limit setting was difficult, as manually pulling and manipulating the data was very cumbersome

Now, Heron Data enables Torpago to lend with much greater confidence, due to:

Wealth of Data: A multitude of powerful calculated metrics and enriched financial data points displayed in Heron

Automated Cash-based P&L: Creation of an automatic cash-based P&L “takes literally a matter of seconds” and enables the team to fully understand and assess the company’s financial position prior to making any credit decisions

Easy and Powerful Portfolio Monitoring: Heron enables tracking of Torpago’s book of business performance over time, enabling dynamic limit setting. This allows Torpago to maximise gains for customers whose finances are improving over time and minimise losses by exiting early for customers whose position is degrading

If you want to expand your addressable market to any business with a bank account, reduce the time it takes you to make a credit / fraud decision, build in portfolio monitoring, or offer a variety of other market-leading money experiences like transaction enrichment, get in touch! We’d love to support. Email sales@herondata.io or book a demo.

Torpago partners with Heron Data to 20x underwriting speed

Until recently Joe Foster was a Senior Credit Risk Manager at Forward Advances (LON: FWD), one of the UK’s leading Revenue Based Lenders. Prior to this he worked at Barclays in SMB Credit Portfolio Management. At Forward, Joe and his team leveraged Heron Data daily when making credit decisions. We sat down with Joe and wanted to share some of his insights on how Heron Data helped him and his team make smart credit decisions in what is the toughest macroeconomic climate most credit professionals have experienced:

“Underwriting SMBs has historically been expensive and time consuming,” said Joe,  “Lenders typically collected bank statements and a set of accounts before calculating ratios and reconciling the bank statements (which are hard to manipulate) with the accounting data (which is not).”

“The result”, as highlighted in the FinRegLab Small Business Spotlight, “is that writing a $100,000 loan costs a traditional financial institution almost the same as writing one for $1,000,000. Since larger loans are typically made to more mature companies they also have lower default rates. These factors worked in tandem to starve newer companies looking for smaller loan sizes of credit.”

“Forward Advances is one of a new wave of lenders who automate their data collection of real time data & reconciliation in order to provide a slick customer experience and reduce operational costs of credit assessment”

“We know that finding reliable financial information on SMEs has historically been challenging. One reason we loved using Heron-enriched Open Banking data is that it made it 100x faster to take all the revenue that passed through the company’s bank account and compare it to the revenue they reported in their P&L. Private Equity funds pay accountancy firms thousands of dollars and  wait weeks for ‘cash proofs’ like this but with Heron we could verify these two disparate data sources against one another in minutes, not days.”

“It didn’t end with revenue - since we primarily lent to small businesses, most don’t have audited accounts. We may or may not agree with how their accountant books revenue, COGS, and other expense items. Heron gave us an unbiased and standardised view into their Gross Profit, Debt Service Coverage, Net Operating Income, and a variety of other metrics. We compared Heron’s output to the numbers which applicants submitted to understand how far they could be out in a worst-case scenario and prioritise applicants for manual review if necessary.”

“Given the current macroeconomic climate, now more than ever you want high confidence in objective information. Even a small inaccuracy in customer-provided information (for example slightly fatter margins or higher revenue than the true number) can affect the outcome of an advance.”

IN TIMES OF UNCERTAINTY, COLLECTIONS START TO MATTER

“‘Revenue Based’ lending gets its name because instead of interest and an agreed amortisation schedule, most companies collect their repayment based on customers’ fluctuating revenue. Most lenders are only able to collect revenue from Platforms with which they can integrate directly (e.g. Stripe, Shopify, WooCommerce). These platforms standardise revenue data, making it easy to know how much revenue a company has generated on a given day.”

“A large part of credit risk is really repayment risk: will the customer take longer than expected to repay? If so, the lender bears the cost because of the fixed fee model. If the customer does go under, how much of the advance is outstanding? The less the better.”

“If you can recognise all revenue that flows into a customers’ bank account, you have a larger pool from which you can collect repayments and you can also make a more competitive offer at the point of application. If a customer generates revenue from B2B sales (imagine you are looking a D2C ecommerce brand that get revenue from Shopify but also retails its products in Tesco), being able to collect on the Tesco revenue in addition to that flowing through the online store can decrease your minimum exposure by a significant amount (many months in some cases). In times of uncertainty, collections start to matter more as the market places a higher value on profitability rather than just growth.”

“In a competitive market, recognising a larger revenue pool also enables you to offer lower interest and higher advance amounts. Heron’s platform made it really easy for us to understand ‘true’ revenue beyond what we could pull from our commerce integrations. They gave us balance analytics that also made it 10x easier for us to make sure we collected repayments at the best time to minimise negative events like bounced direct debits”.

UNDERSTANDING CUSTOMERS' BORROWING STACK IS CRITICAL

“Revenue based financing is a competitive market, there are multiple players and lots of customers have relationships with other lenders. Understanding who else was in our lending stack, and being able to easily estimate the amount outstanding to each player, made it far easier for us to assess the true risk presented by an applicant. There were multiple cases where the only evidence of a customers’ indebtedness was their bank data because the loans weren’t disclosed by the applicant upfront or present in the other financial statements we collected."

“In the current climate, insolvency events are more likely than ever. Since pretty much all revenue based lending is unsecured, understanding customers’ borrowing stack is critical. I want to know how many other players were collecting against the same revenue and who else had how much to lose. Heron make this really easy.


Heron Data supports Fintechs and B2B lenders like Forward Advances, ClearCo,Outfund, Settle, Pipe, and 85+ others to automate loan decisioning with bank transaction data. If you’d like to learn more or are interested in testing us out email sales@herondata.io or book a demo.


Three ways Heron Data helped me make better lending decisions in a downturn

Bank transaction data is one of the most widely available and tamper-proof data sources on any company. It's easy to share, and always up to date. However, extracting quick insights from this data is difficult. Unlike accounting data, bank data data is unstructured and not normalised. Heron Data helps over 65 lenders and fintechs from 12 countries understand and evaluate businesses using bank transaction data. We have enabled customers to speed up their manual underwriting process by 8x and have helped them avoid giving out millions of dollars in loans to high-debt or otherwise risky applicants.

Today, we're excited to launch a new feature that will make it even easier to extract these types of money-saving insights from bank data: Heron's new scorecard. Instead of calculating features from categorised data, the scorecard enables our customers to quickly see a range of insights. For example:  

Completeness and quality of data: Unconnected accounts, range & density of transaction data

Risk flags: (Lack of) tax payments, large ATM withdrawals, insufficient funds or overdraft fees

Debt metrics: Size and recency of debt, number of lenders, debt service coverage ratio, debt repayments

Traditional finance metrics: Change in closing cash balance, Cashflow from Operations, Revenue collected, etc.

The scorecard can either be seen in Heron's dashboard, or consumed via API to power underwriting models or be shown back to underwriters in their normal underwriting environment.

To give each metric context, we're also showing you how each metric compares to all companies across your portfolio. On an opt-in basis, you can also see how it compares to the combined portfolio of all other Heron customers who have opted-in. For example, you can see whether the data quality of a company is worse or better than average or whether its debt service coverage ratio is above or below 90th percentile across your entire portfolio, etc.

To see the scorecard in action, you can watch the video below. To find out more, you can either sign up for a free trial, or book a demo!

Heron Data Scorecard

Understand and benchmark a company in seconds with Heron's Scorecard

The security and privacy of our customers' data has been at the forefront of our minds and decisions since we started. As a bank transaction categorization and enrichment service, we deal with sensitive data. That's why we have always had strong measures and practices in place to ensure this information is kept safe.

In order to further validate our commitment to security, we underwent the process of being certified as SOC 2 Type 2 compliant. This certificate is the most accepted security compliance standard so we wanted to give our customers peace of mind that we follow the latest best practices around security both from a technology but also internal processes standpoints.

As a SOC 2 compliant company, we take important measures to ensure our customers' data is always safe such as:

Encryption of data in transit (from our customers to us).

Encryption of data at rest (when it sits in our database).

IP-restricted access to our internal services, databases and configuration.

Monitoring access to and configuration of our systems, including automated alerts in case of anomalies.

Monitoring of employee work computers to ensure we're all using secure processes, such as short computer lock times, use of a password manager, hard disk encryption and antivirus use.

Thorough review of our vendors to ensure they're all SOC 2 compliant.

Periodic penetration testing of our systems conducted by a third-party.

We encourage you to have a look at our security and disclosure pages for more information on our security practices.

If you are an existing or prospective customer and would like access to our full SOC 2 Type 2 report, please get in touch at security@herondata.io and we'd be happy to provide it.

SOC 2 Type 2 Certified Transaction Categorization and Enrichment

We know that integrating a transaction enrichment API can take time, especially when you have limited tech resources. We also know that sometimes API integrations don't work for every bank and so customers or intermediaries can end up sharing CSVs which are time-consuming to review.

Our customers can now manually upload CSV files in our dashboard in both of these cases. For any company, click the Transactions dropdown, select the desired upload method, follow the steps and let our industry-leading ML do the rest.

You can upload multiple files for the same company if you want to add additional transactions over time and we'll do all the work to provide you with a single, unified view.

If you are a lender working with bank data and would like to test Heron out, book a call. We'd love to show you how we can add value to your organization with a free trial.

Start analyzing bank data without writing a single line of code

Although bank data is the single source of truth of what is going on inside a business, it can be messy and hard to work with in its raw form.

Heron makes bank data far easier to work with by categorizing and enriching unstructured data into a set of metrics representing a company's financial health. For example, our Machine Learning models allow you to recreate a Profit & Loss statement in real-time:

Today we are pleased to announce the launch of our underwriting dashboard. The dashboard is a tool Heron customers can use to interact with companies they have sent to our platform. The dashboard both visualizes company's data to give you a quick sense of an applicants' financial health and allows underwriters to deep-dive and re-categorize transactions where they might disagree with an assigned category.

If you are a lender working with bank data and would like to test us out, book a call. We'd love to show you how we can add value to your organization with a free trial.

Visualize a company's financial health with our new dashboard

As a startup trying out a new tech idea, Heroku is a great place to start. You can deploy your application code easily and all your dependencies are managed for you. However, once you start seeing real throughput, moving your dynos to Kubernetes makes a lot of sense. Here we explain how to easily imitate a common Heroku setup in Kubernetes.

Let's first define what a Heroku infrastructure is based on: web and worker dynos.

Dynos are a unit of scaling in Heroku. Each one is a virtualized and isolated Linux container. To scale your application to deal with more traffic, you can increase the number of dynos you're running (either manually or automatically). These are also ephemeral, meaning they can be shut down and restarted as needed.

Usually you can split your dynos into two types: web and worker. Web dynos deal with HTTP traffic and workers are used for asynchronous computation which can be more time consuming than what the scope of a 30 second web request allows. You can scale web and worker dynos individually.

So to summarize what you need to have in Kubernetes in order to achieve feature-parity:

A container-based way of running one program at once

Ephemeral containers which can be shut down at short notice

A way to distinguish types of containers, and scale them separately

A way to scale the number of each type of containers

In K8s, the closest way to mirror a Heroku web and worker setup is to use deployments and pods.

A pod is defined as "the smallest deployable units of computing that you can create and manage in Kubernetes". This is basically the equivalent of a Heroku dyno. You'll want to have one pod definition for web dynos, and one for worker dynos. You can then define their replication and scaling using deployments. Again, one for web pods, and one for worker pods.

With deployments, you can manually set the number of pod replicas you want, or you could link the deployment to a horizontal pod autoscaler for automatic scaling based on CPU usage. With a horizontal pod autoscaler you can define the minimum and maximum number of pods for a particular deployment type, and Kubernetes will handle the rest.

One of the trickiest decisions while migrating over is how much CPU and Memory each pod should have allocated. In Heroku, you can choose a dyno type and you're done. In Kubernetes, you have more control but also more configuration to set up. Our approach was to configure resource requests and limits per pod which were roughly equivalent to our dyno types in Heroku. From there, once your Kubernetes system is live you can tweak if needed.

Here we've outlined how to directly copy a Heroku infrastructure into Kubernetes. While this can be helpful for an initial systems migration, it is almost certainly not ideal. Systems designed for Heroku will likely not be optimized to run in Kubernetes and take full advantage of its features. So we encourage you to give this a try, monitor your systems metrics and tweak as needed.

We hope that you found this helpful! Thank you for reading.

How to replicate a Heroku infrastructure in Kubernetes

‍TLDR: make your transaction feed delightful with our free merchant extraction endpoint.

Whether you work with card data from Marqueta, bank data from Yodlee / Plaid, Open Banking data from TrueLayer, or Direct Deposit Account data from Unit you have probably spent cycles cleaning and normalising the data. We certainly did.

When building our first Fintech we spent tons of time writing regexes, populating a merchant database, building NLP, NER, and deep learning models for categorisation, identifying recurring transactions, and much more. During Y Combinator, so many companies asked to use our tech that we decided to become Heron Data.

We power transaction enrichment for Business- and Consumer- focused Fintechs and FIs so they can spend more time shipping features

Our customers love us because we make it 100x faster to:

Engage users with a beautiful transaction feed complete with with clean merchant names and icons. Our payment_processor field even makes dealing with pesky intermediaries like PayPal a breeze.

Shed light on spend with powerful merchant- and category-analytics dashboards

Increase cards issued per customer by offering card merchant locking: allow your users to select the merchant they want to blacklist / whitelist from our merchant search endpoint, post the transaction to Heron in the Auth flow, and fail transactions without the correct merchant_heron_id

Speed up accounting 4x by writing rules based on Heron merchants and categories rather than raw transaction descriptions

Now, we enrich $10bn+ in transactions every week and over 1m individual transactions each day. For our typical customer, we categorise 100% of transactions with 94.6% accuracy. We extract the merchant from over 80% of transactions with a sub 1% error rate and return a human readable string where we can't enrich the merchant with metadata.

Today we are excited to launch our free merchant extraction endpoint so you can focus on building, not making your transactions look nice.

‍Book a demo if you need a higher rate limit or are interested in our underwriting solutions, categories, recurrence, or other capabilities.

Heron Data launches free merchant extraction API

We're delighted to announce Heron Data's first funding round, a $1.2M pre-seed round completed in September. We're excited to be backed by a slate of great investors, including Y Combinator, BoxGroup (Plaid, Stripe), Flex Capital (Chime, Flexport), Musha Ventures (Streak, Lever) and a range of angels, including Fintech unicorn founders Matt Robinson (GoCardless), Shivaas Gulati (Remitly) and Jonathan Levin (Chainalysis). We have raised this round to help us move faster towards our immediate goal, which is solving the problem of understanding and enriching bank transaction data.

You may sometimes look at your bank statement and ask yourself what a transaction description like the below could possibly mean:

SP * PINEAPPLE CO 05/06 PURCHASE 098314763

As infrastructure allowing third-parties to access bank transaction data is growing around the world, more and more companies have to know the answer to that question, too. A new class of start-ups is building anything from budgeting apps to tax preparation software on this data. Banks that do not enrich and act on this data risk being left behind by the start-ups that do.

In building an earlier company that was leveraging such data, we realised that the problem of making sense of this data by labelling and categorising it is unsolved. Most companies rely on in-house solutions that are constrained by small data lakes, mapping existing category labels to the custom categories that their business logic requires.

We are solving this problem today for multiple paying customers, and you can get a sense of our approach here. We're offering a categorisation service with bespoke labels, in which our customers benefit from a data lake much larger than their own. The impact of a full suite of world-class products on transaction enrichment are vast: Think of the hours spent by book keepers reconciling bank transactions to accounting entries; or the resources needed to reply to consumers disputing legitimate charges on their bank statement that they cannot identify.

As for the above, it turns out that the transaction comes from the Pineapple Collaborative, a company with the slogan “pantry staples, made by women”, making Olive Oil locally in California. Depending on your use case, this may be tagged as “Groceries” or "Supplies", but it could also be "Personal expense", "Independent retailer", or "Low-carbon retail".

Solving the problem of labelling such an object using data science models (as opposed to a tagging farm) poses interesting technical challenges. This is because our object of analysis is unique: A transaction description string, an amount and a timestamp. This means that:

We can use elements of the advances in NLP (like GPT-3), but we are not quite analysing natural language, so these do not just work off-the-shelf.

We operate in an environment without a ground-truth (e.g., multiple underwriters at a Fintech lender may classify the same transaction differently), so we need to build technology that weighs the judgements of different annotators to come to a consensus label for a given transaction

We can calculate features of the data that are powerful in predicting category labels, for example the recurrence of a transaction

More broadly, we believe that the "Finance edition" of software-is-eating-the-world is barely underway. By the time it is finished, every financial institution will be a software company. We will enable that change by building technology that makes it easier to create intelligent, personalised digital financial products with minimal overhead. With advances in NLP and machine learning, a future where every consumer and SME can profit from the kind of bespoke financial advice and services currently reserved for the rich and large corporations is within reach. We want to power that future.

If raw transaction data is causing your business pain, email us or book a demo on our website - we'd love to work with you.

If you want to join us, have a look at our Careers page - even if no position is open, we always enjoy talking to people interested in what we do.

If you want to be kept up to date on our progress, drop us a line at hello@herondata.io, and we'll send you semi-regular updates.

Heron raises $1.2m to help FinTechs make sense of bank data

Heron Data Blog