Employee Spotlight: Interview With Matthieu Oung, Financial Modeling and Product Analyst at Eqvista
Some people stumble into finance by default – Matthieu Oung chose it on purpose. After a general first year at Audencia Business School, he deliberately went after corporate finance, with a particular pull toward startups and venture capital. It’s a path that has taken him from inside a French SaaS startup during its own sale process to his current role as Financial Modeling & Product Analyst at Eqvista.
What makes Matthieu’s perspective worth listening to is the thread that runs through everything he does. Early on, he noticed that where founders are most exposed, they see their product clearly, but not their own finances, which is exactly where the big decisions get made. That observation has shaped his entire career. For him, it all comes down to translation: turning numbers into a story someone can act on.
At Eqvista, that means building valuation models that don’t just work on paper, they have to hold up when someone he’s never met uses them, on their own data, without him in the room. It’s a high bar, and it’s one he seems to genuinely enjoy clearing.
We sat down with Matthieu to talk about what drew him to the startup world, what it’s really like to build financial tools from the ground up, and why he believes an elegant but unused model is just a hobby. Here’s the full conversation.
Matthieu, every finance career has an origin story, what was yours? What pulled you toward numbers and models in the first place?
I did a general first year in Audencia Business School, so corporate finance was a choice rather than a default. I became interested in corporate finance, especially targeting startups and venture capital. Getting hands on the startup environment and its financial side was clear to me early. I noticed where founders are most exposed: they see their product clearly, but not their own finances, which is exactly where the big decisions get made. So I went after both sides. As a finance manager and FP&A analyst, from the inside. As a valuation analyst, from the outside. Both come down to translation: turning numbers into a story someone can act on. That is what pulled me in. I have always wanted to know why something happened.
Was there a specific moment when startups clicked for you, a project, a company, a conversation, that made you think, “this is the world I want to work in”?
I was inside a French SaaS startup while it went through its own sale process: the vendor due diligence, then the sell-side work. In parallel, I helped my manager build the business plan for a fundraising. So I was not advising from the outside. I was in the building, watching the daily operations of the business at the same time as the event that would decide its future. That is what a startup gives you that a large company does not. You see the whole thing at once: the operations, the business model, and the numbers meant to describe both. You also see every party at the table wanting something different from the same figures. Those two projects are the foundation of everything I know in finance. Contributing to something that is consequential as a junior is rare, and it settled the question of where I wanted to work.
What were some of the biggest early challenges you faced when moving from finance and FP&A into a product-focused role?
In FP&A, a model is finished when it is correct, when the figures match one another across different sources, and when you can defend it in a meeting. In product and financial modeling, a model is finished when someone you have never met can use it, on their own data, without you in the room. That is a much harder bar. The second challenge was edge cases. When I built models for one company, I knew that company: the business model, the suppliers, the teams. At Eqvista, a model ships into the platform and runs across thousands of companies with messy data, missing fields and unusual structures. Suddenly the assumption you made without thinking becomes a bug. That is why you have to anticipate the full range of cases upfront, to reduce the risk of errors in the model.
Everyone has that one mentor, project, or early experience that quietly shapes how they work years later. What’s yours?
Real-Time Valuation. I was involved from the beginning, and I will admit my first reaction was that it could not be done. Pricing a private company continuously and charting it the way a public stock trades: I did not see how. Private companies do not trade. There is no tape to read. But the lesson I took from it is not a technical one. It took the analyst team, the IT team, the product team and the sales team, and there was not a step you could remove. The methodology was worth nothing until it was implemented. The implementation was worth nothing until someone made it legible to a customer. And none of it mattered until the people speaking to customers told us what they actually needed to see. We are spread across the world and rarely in the same time zone. What made it work was that everyone understood the same goal. That is what stayed with me: alignment matters more than proximity. I owe that team a lot. None of it would exist without them.
Let’s talk about what it’s actually like working at Eqvista. What’s the culture like on the inside, how do people collaborate, push back on ideas, or support each other when things get tough?
I have worked directly with the teams in Hong Kong and in Europe, and what strikes me is how short the distance is between a question and a decision. You raise something, you get a response, it gets acted on. Very little time is spent restating a problem everyone already understands. That works because people have genuine expertise in their own domain. When I bring a valuation question to an engineer, or an engineer brings me a data problem, neither of us has to explain the basics first. You can be direct and technical immediately, which is faster and more honest. The other thing is precision. In this business, a number that is roughly right is not right, and everyone here treats it that way.
For those unfamiliar with your work, how would you describe what a Financial Modeling & Product Analyst does at Eqvista?
It is two jobs held together. The first is the finance one: deciding how a private company should be valued, which comparables, which multiple, which assumptions actually carry the result, and where the method breaks down. That part looks like valuation work anywhere. The second is turning it into something that runs on its own. A model I build for one company is an argument I can defend in a meeting, but a model inside the platform has to hold up across thousands of companies, on data I have never seen, with nobody in the room to explain it. So I write the specification, work through implementation with the engineering team, then test the output against my own model to find where the two diverge. And because a number nobody understands is worth nothing, part of the job is how it is presented: the chart, the context around it, what a founder sees first. Concretely that means real-time valuation, valuation reports, and the financial modeling features across the platform. But the thread is the same as it has always been, translating numbers into something someone can act on.
What does a typical week look like for you when you’re building valuation models and working with the engineering team?
The rhythm is set by a weekly review with the product team. That is where work gets checked, challenged and pushed forward, and no step moves to the next one without that backing. Nothing here is carried alone. Around that review, the week has three modes. Part of it is analytical: building or refining a model, pulling and validating market data, testing how a change in methodology moves results across a whole population of companies rather than a single case. That is mostly Excel and research. Part of it is specification and review with the engineering team: walking through the logic, agreeing how a feature should behave in the ordinary case and in the awkward ones, then reconciling the implemented output against my own model to find where the two diverge. That reconciliation is where the real problems surface. A difference of a few percent is usually a convention, but an order of magnitude means something structural is wrong upstream, and you only find it by checking. And part of it is product: how a number is displayed, what context sits around it, whether a founder understands what they are looking at without an explanation.
What part of your role feels most rewarding: solving complex financial problems, shaping the product experience, or seeing the platform help customers?
Seeing it land with customers, though I will be honest about why. Solving a hard financial problem is satisfying in a private way. Shaping the product experience is satisfying in a craft way. But neither means much until someone uses the output to make a decision. A model that is elegant and unused is a hobby. The moment I look for is when a founder looks at a valuation and it changes what they were about to do. That is the whole point of building it inside a product rather than delivering a file.
From your perspective, what makes a strong valuation or equity management product stand out in a crowded market?
Being auditable rather than authoritative. A lot of tools give you a number. The ones that earn trust show you where the number came from: which comparables, which multiple, which assumptions, and how sensitive the result is to each. Founders are smart, and they are going to be challenged on that number by an investor or a board. If they cannot defend it, the product has failed them even if the math is right. The second thing is coverage of the exceptional cases. Any tool can value a clean SaaS company with three years of revenue. The differentiator is what happens with a pre-revenue company, an unusual capital structure, or a stale last round.
What do you think about product-market fit when building features for private companies with very different needs and growth stages?
You cannot build a different product per stage, and you should not try to build one that averages them. What works is a single engine with different anchors. An early-stage company with limited revenue history has to lean more on market comparables and the structure of its last round. A growth-stage company with real financials can lean on its own numbers. The methodology is shared. What changes is the weight you give to each input. The mistake is treating stage as a cosmetic difference, with different copy on the screen and the same logic underneath. Stage genuinely changes what evidence is reliable.
What signals tell you that a financial modeling feature is truly useful for customers, rather than just technically impressive?
First, people use it without being told to. If it needs an explanation every time, the feature is not done. Second, they argue about it. A user who challenges an output is a user who takes it seriously, and indifference is much worse than disagreement. Third, and the strongest: they take it somewhere else. When a founder exports the output and puts it in front of an investor or their board, they have staked their own credibility on it. Nothing technically impressive earns that.
What has been one of the hardest challenges in building valuation models for private companies, and how did you work through it?
The data is a major challenge while building models. With a public company the price is handed to you. With a private company you must go and find it, and most of the time it does not exist in any one place. It is spread across specific databases, filings, funding announcements and secondary marketplaces, so a lot of the work is plain research before any modelling starts. Even then you will never reconstruct a price with the precision a public tape gives you. Two defensible methods applied to the same company will not land on the same number, and pretending otherwise would be dishonest. You see it most clearly when prices disagree: a primary round price from the last fundraise and a secondary price from shares changing hands often do not match, because they reflect different buyers, different rights and different liquidity. Neither one is wrong. Working through it meant separating what is explained by structure, things like preferences, illiquidity and the gap between preferred and common, from what is genuine information about the company’s value. Once you can do that, you can decide how much weight each source deserves instead of picking one and hoping. So the challenge is not finding the right number. It is combining incomplete information carefully and then being explicit about what went into it when you display the result. With private companies, transparency is what makes a valuation defensible.
Working with financial data and product logic can be complex. What’s a lesson you’ve learned about simplifying complexity without losing rigor?
Simplifying is not the same as making the model simpler. That is the shortcut, and it costs you the accuracy you built. It starts earlier, with being clear about what the model is for and which question its output has to answer. Once that is settled, you know what has to be visible and what can sit underneath. Then keeping the rigour is a matter of understanding your own model well enough to make it legible, to a finance expert and equally to someone who has never opened one. I think of it as translation. Some sentences cannot be carried word for word into another language. If you try, you get something technically faithful and meaningless. You have to find different vocabulary that expresses the same thing to the person reading it. A model is the same. You are not lowering the standard, you are changing the words, and the standard is whether the reader ends up understanding what you understand.
How do you see financial modeling and valuation tools evolving over the next few years for startups and private companies?
Two shifts, I think. The first is from point-in-time to continuous. A valuation has historically been an event. You commission one, you get a document, it is stale within months. There is no good reason for that anymore. Market data, comparable company performance and a company’s own numbers all move constantly, and the valuation should move with them. The second is data coverage. The gap between public and private markets has always been an information gap, and it is closing: more secondary market activity, more structured filings, more accessible data. That makes it possible to value private companies with something much closer to the rigor applied to public ones. AI accelerates both, mostly by making unstructured information usable. But it does not replace the methodology. It gets you better inputs faster. Deciding what counts as a reliable input is still the hard part, and still a judgment call.
If you could build one “dream feature” for startup finance teams, what would it be and why?
Coming from FP&A, I would build the bridge between a company’s own numbers and its valuation. Today most models value a business from the outside: comparables, transactions, what the market pays for companies that look like yours. That is necessary, but it is half the picture. The other half sits inside the company: the monthly, quarterly and yearly reports, the budget, the cash flow forecast, the investment plan. Those are the things a founder controls. So I would want a founder to be able to ask what 10% revenue growth a year does to their business, and what it does to their valuation. The same for a hiring plan, a change in burn, an investment delayed by two quarters. Right now those two questions live in different tools and get answered by different people. Put them together and it stops being a reporting tool. A founder sees their own operations and how the outside world will read them, at the same time. That is the decision tool for the whole life cycle of a company.
