Peter Lord and the trouble with moving a number

A number can travel around the world in a fraction of a second and still take weeks to get from a business’s accounts to its bank. Peter Lord built a company in that peculiar interval. The money had been earned, the invoice recorded, the information stored. Somehow, somebody still had to send an attachment.

Lord encountered the problem while working at MarketInvoice, the British invoice-finance business later known as MarketFinance. He built connections to bookkeeping software, then moved into product management and inherited the list of connections still waiting to be built. There were always more systems. Other companies would need them too. A recurring engineering chore began to look like a business.

In 2017 he started Codat with Alex Cardona and David Hoare, colleagues from MarketInvoice. Almost a decade later, he remains its co-founder and CEO. The company has moved from supplying financial-data connections toward helping commercial banks interpret what comes through them. Along the way, Lord has had to answer an awkward founder’s question: when customers keep asking for more, which part should you actually build?

The list that became a company

His route into software included a change of course at the University of Southampton, where he began with geography before switching to computer science. His public LinkedIn profile records first-class honours and attendance from 2010 to 2013. Credit Suisse also appears in his early career, before his engineering and product work at MarketInvoice.

That sequence matters because Codat grew out of a job he had already done. A lender needed information that its customers’ existing software already contained. Connecting directly could spare a business the exercise of exporting and supplying it again. Having built those links himself, Lord understood the work concealed inside the deceptively small word “integration”.

As product manager, he could see the queue stretching ahead. Building another connector helped customers using that particular package. It left the next package waiting. Selling that work as a shared service offered a different way to tackle the queue: several financial providers could use the same underlying connections.

There is a satisfying lack of theatre in this origin. No customer needed a new appetite for accounting. Financial providers already needed the records. Lord and his colleagues had lived through the difficulty of obtaining them. The opportunity was in repeating the useful engineering once, then making it available to others.

A chief executive with code to write

In Codat’s first year, Lord described a CEO’s responsibilities through the product. Roughly half his time went into coding; the rest went into meeting clients and their users. The business was essentially bootstrapped at that point. Its ambition was specific: a developer should be able to connect through Codat in less than a week.

Asked which phone he carried and why, he supplied a miniature rebellion against the technology interview: My phone. Because it’s mine. The answer is funnier for its refusal to help. A founder selling connections had discovered a question to which he preferred a dead end.

The humour sits beside a fairly consistent product argument. Codat would standardize information coming from different accounting platforms. A financial provider could work with a common format rather than start again for each package. The small business could keep using the software it had chosen.

A common format sounds tidy. The work beneath it is less obliging. Software packages organize records differently, and accounting conventions change across borders. An interface can look simple precisely because somebody else has spent time accommodating those differences. Lord’s early proposition put that somebody else inside Codat.

The customer who may never know your name

The person doing the bookkeeping was often a step removed from the sale. Codat sold to financial institutions and software companies serving businesses. That arrangement required Lord to keep two constituencies in view: the organization buying the infrastructure and the business whose information made it useful.

A concrete example appeared in 2021. Zettle, the PayPal payments business, used Codat to transfer point-of-sale transactions into merchants’ accounting software. The benefit lived in the merchant’s working day. Sales information could reach the books without another round of manual entry.

That March, Codat announced investments from PayPal Ventures and American Express Ventures and formally launched its US operations. Index Ventures had supplied $10 million in funding the previous June. These were commercial relationships with firms that had their own reasons to care about the same difficult journeys between systems.

By July 2022, Codat had more than 200 corporate clients. Its expansion brought another reminder of what “universal” costs in practice. Lord described work on a German bookkeeping integration and the associated differences in language, accounts and tax. Taking the company abroad meant adapting to how people actually kept their books.

His example of the desired experience was pleasantly ordinary: a restaurant booking appearing automatically in a calendar. Consumer software had trained people to expect that small convenience. He wanted business software to provide similar cooperation. A restaurant reservation is a modest comparison for a financial-data company, which is part of its usefulness.

Forty-two ways to lose your focus

In an October 2022 conversation, Lord counted 42 uses for Codat among its clients. They ranged from payroll reconciliation to corporate cards. For infrastructure, that range creates opportunities and a management problem. He put the problem plainly: where do we focus?

A product serving many markets can receive a very persuasive request from every direction. Lord named expense management, point of sale, banks and lenders as areas of concentration at that time. The company could improve shared infrastructure while choosing where to spend more attention. Every possible application could not become an equal priority.

“where do we focus?”

Peter Lord, discussing Codat’s range of uses in 2022

That same year brought a $100 million Series C financing, announced in June. J.P. Morgan, Canapi Ventures, Shopify and Plaid were among the investors. The funding made a conspicuous headline. Lord’s recurring problem remained smaller and more practical: deciding which work would make the service more useful.

Growth also changed the work of leadership. In a 2021 interview, he identified hiring the right people as a major challenge. The early CEO could spend half his time writing code. A company serving more clients needed other people making good decisions too. A founder’s familiarity with a problem is valuable; it does not fill every vacancy.

First the connection. Then the calculation.

Lord’s account of launching Assess in 2022 reveals a deliberate hesitation. Customers had asked Codat to add analysis for years. The company had resisted while concentrating on reliable data infrastructure. Only after processing millions of financial datasets did it add the intelligence layer.

Assess offered lenders prepared metrics from accounting, commerce and banking information, including financial ratios and categorized transactions. Lord said 25 finance providers were already using it when he announced the product. Codat was moving further into the customer’s decision process: obtaining a record, then helping extract something useful from it.

The distinction helps explain the later company. A bank can receive a large amount of information and still have substantial work ahead. Someone must decide which entries belong together, which patterns deserve attention and which figures matter to a particular conversation. Getting the information through the door leaves plenty of work inside the room.

In April 2024, Codat announced Spend Insights, already in production with J.P. Morgan. It used connections to accounting and enterprise resource planning systems to provide supplier and spending data for virtual-card programs. Persistent connections replaced static files, making the exercise repeatable as the underlying records changed.

This was a more particular task than the original promise of easy integration. Which suppliers might be approached about accepting a different payment method? What spending was relevant? The connection remained essential, while the product’s purpose became easier to describe in the language of a bank’s daily work.

A different conversation at the bank

In May 2026, Codat announced an advisory intelligence platform and a strategic repositioning toward commercial and business banks. BMO was among the first banks collaborating on the new offering, using Spend Insights to support middle-market commercial-card conversations. Codat reported having powered more than 350,000 connections to business customers’ financial systems.

The vocabulary had moved some distance from the early API pitch. The underlying question was still recognizable. A business had information in its systems. Its bank needed a better understanding of what was happening. Codat proposed bringing the two closer together, then turning the records into something a banker could discuss.

The 2026 brand refresh acknowledged that the company’s identity had lagged behind its products. Even the name retained a trace of the beginning: Codat comes from company data. The new positioning emphasized interpretation alongside connectivity, with commercial banking giving that interpretation a defined audience.

For Lord, the interesting test is how much of the original irritation can disappear. He has imagined a future in which integrations are so routine they cease to be a marketable feature. Until then, there are books to connect, figures to understand and customers waiting for an answer. A number has a surprisingly busy journey. He would like less of it to involve an attachment.