One revealing sentence Ben McRedmond has written about building software is only eight words long: “Well, it helped that I failed once before.” It sits in the story of Equals, the company he now runs, and it carries more weight than the usual founder folklore about garages, genius, and inevitable victory. McRedmond’s career is interesting because the second company does not erase the first. It uses the first as load-bearing material.
Equals is an AI analytics platform shaped like the tool business people already reach for when the official dashboard cannot answer a follow-up question: a spreadsheet. It connects live company data, lets teams query and model it, and turns recurring questions about revenue, retention, pipeline, and customer behavior into workbooks that can keep updating. The current pitch is about trusted answers. The deeper product idea is about familiarity. McRedmond and co-founder Bobby Pinero decided that the cells, formulas, pivots, and free-form canvas were fundamentally right. The pipes beneath them were due for reconstruction.
This conviction arrived by an indirect route. McRedmond did not attend college. He has said that his teenage education consisted of obsessively reading about software and how companies were built. There was theory in that education, certainly, but no graduation ceremony suggested the learning was finished. His advice to new graduates is similarly practical: if you want to build things, start by building them.
First employee, many jobs
In 2012, McRedmond joined Intercom as its first employee. He began as an engineer, the useful generalist in a company where job boundaries had not yet hardened. Intercom co-founder Eoghan McCabe later described the logic with cheerful simplicity: they were a software company, so they hired an engineer. More important, McRedmond was ready to pitch in wherever the young company needed him.
That generalism became a career. McRedmond eventually founded Intercom’s growth team and argued for a definition that avoided the usual bag of conversion tricks. A growth team, in his formulation, was a software team solving a company’s business problems. The distinction mattered. Product teams worked on customer problems. Growth teams applied similar craft to how the company reached, converted, and served those customers.
His favorite example was a tiny coffee shop down an alley near Intercom’s San Francisco office. The place had an espresso machine and little else. It became his parable for premature optimization. A growth team could polish the cash register endlessly, but the shop would do better to improve the space, keep customers dry, and make the entire experience worth entering. Measurement was useful, but it came at a cost. The measurable thing was not automatically the important thing.
The argument sounds obvious once stated, which is one of McRedmond’s recurring tricks. He wanders outside software, notices an old human activity hiding beneath a new technical label, and returns with the pretension removed. Onboarding resembles an instruction manual. Product discovery can learn from a bookstore. A Roman seller floating food in water was practicing growth hacking centuries before anyone put the title in a bio.
The email company with too many ideas
McRedmond left Intercom in 2017 to co-found Consider, an email client for work. The premise was humane. Slack had made workplace communication immediate and expressive, but it also allowed interruption to become ambient. Email was asynchronous and organized by topic, qualities that Consider wanted to restore without preserving every stale ritual of the old inbox.
The launch produced a waitlist of thousands overnight. Then the product met the ordinary cruelty of use. Consider had changed how conversations worked, how the inbox behaved, and how messages appeared. Meanwhile, familiar essentials such as spellcheck, labels, and snooze were missing. Its meaningful ideas were obscured by the effort required to operate it. After several years with little traction, the company shut down in 2020.
McRedmond’s post-mortem did not blame timing or timid customers. He thought planning had failed, and he thought the team had introduced too many independent risks. The brutal insight was mathematical in spirit: a product’s chances were not the sum of its bets but the product of them. Multiply enough uncertain assumptions together and one bad answer can drag the whole thing toward zero.
Where the uncertainty lived
Conceptual map based on McRedmond’s account: Consider changed several core conventions at once. Equals rebuilt a large foundation but concentrated novelty in live data connections.
This was not a lesson to build less. Email clients and spreadsheets share an inconvenient property: neither becomes useful after a weekend of tasteful prototyping. People arrive with decades of expectations. They do not congratulate a new spreadsheet for rendering a cell. They notice when paste is odd. They notice when the formula grammar is incomplete. They notice everything.
The case for the whole thing
When McRedmond and Pinero began Equals in 2021, skeptics told them a browser-based spreadsheet was too complicated for a small team. McRedmond agreed with the first half. The product required a programming language, a virtually rendered infinite sheet, an editor, charts, pivot tables, formatting, and real-time collaboration. The useful disagreement was over what complexity demanded.
His answer was the “Full Build”: map the complete foundation, change nothing that did not need changing, and win with one large bet. For Equals, that bet was data connections. The spreadsheet would look and behave like a spreadsheet. Its new power would come from pulling live information out of databases, warehouses, and business applications instead of trapping people in a swamp of copied CSV files.
The apparent paradox is the good part. Equals was technically ambitious and behaviorally conservative. It built a great deal so the customer would have to learn very little. This separates complexity from novelty, two words that startup advice often bundles together and labels “scope.” McRedmond’s doctrine permits a vast amount of work while rationing uncertainty.
It also explains his affection for thorough software from the 1990s. Those products could be slow or homely, but they were complete. Their settings had settings. Modern software learned to find narrow markets and ship smaller products, an economic advance that sometimes became an aesthetic of incompleteness. Equals borrows the older ambition without asking users to abandon the grammar they know.
Judgment before the decimal point
McRedmond is equally suspicious of precision performed before a company has earned it. He has argued that A/B testing is usually too slow for startups and can reduce tangled strategic questions to one falsely tidy variable. An early company may spend a month producing one statistically defensible answer when experienced judgment, customer conversations, and several imperfect decisions would move it farther.
This is not rebellion against numbers from the CEO of an analytics company. It is a demand that numbers answer questions they are equipped to answer. Founders gradually build a model of the business through accumulated contact with customers, product behavior, sales calls, and mistakes. Quantitative movement becomes another signal inside that model, not a mechanical replacement for it.
Equals itself provided the demonstration. For four years the company described its work as building a better spreadsheet. The phrase accurately named the technology and poorly named the reason to buy it. Finance and revenue operations leaders were not shopping for a prettier grid. They wanted to understand how the business made money from first touch through renewal. In 2025, McRedmond said the company had begun framing Equals around GTM analytics. The product stayed flexible; the market story became concrete.
From product chief to company chief
In December 2024, McRedmond took over as CEO while Pinero became Executive Chairman. The change was presented as an evolution of a partnership, not a palace drama. The two had worked together in Intercom’s early years, when Pinero analyzed the business and McRedmond built product flows around its funnel. They knew each other’s work before they shared a cap table.
The new role forced another handoff. After four years running product at Equals, McRedmond transferred day-to-day responsibility to the company’s first dedicated product manager in 2025. It was a small organizational milestone with a large emotional undertow: the founder who had written the specifications would now have to build the conditions in which somebody else could write them.
Meanwhile, his view of AI remained consistent with the rest of his product taste. He dislikes the sparkle button that summons a text box and announces its intelligence. AI, he has written, should be a substrate for features: quietly debugging an error, supplying context, or offering help at the moment it becomes useful. The technology becomes more useful when the interface stops showing it off.
A disciplined second act
There is no clean moral in building companies, only better constraints. Consider’s failure did not prove that email was impossible or that novelty was foolish. It gave McRedmond a more discriminating question: which part of this product truly needs to be new? At Equals, the answer moved from the visible grid to the living data beneath it, and more recently from selling the grid to selling the business answer.
McRedmond’s transferable idea is therefore not “build a spreadsheet.” It is to respect the accumulated intelligence inside familiar tools. Make a full map when the product requires one. Do not mistake a small team for a small undertaking. Keep the speculative surface area narrow. And when a failure offers a specific lesson, resist sanding it into an inspirational quote.
The spreadsheet is an unglamorous object with a peculiar immortality. Its blankness tolerates the user’s imagination. McRedmond is betting that the right successor will preserve that freedom while making the answers fresher, more collaborative, and harder to get wrong. It is a second act built in cells, but the formula began in an inbox that did not work.