Guest: Matt Vega & Sidharth Shah
Today I am talking about fintech fraud strategy, and specifically what it looks like when the business model itself makes risk harder to read. Because that is usually where things get messy. Not when the fraud is obvious, but when the customer base is legitimate, fast-moving, and operating in ways that do not fit neatly into traditional verification models.
I sat down with Matt Vega and Sidharth Shah from Novo to talk about what it took to build fraud operations for fintech around a customer base with unique behaviors, limited traditional data, and a lot of pressure to get decisions right without slowing down the business. And honestly, that is where a lot of fraud teams live. Somewhere between incomplete signals, competing priorities, and the reality that the wrong control in the wrong place can create a completely different problem.
This conversation is really about business model fraud complexity. What happens when standard assumptions break down. What happens when your customers are startups and SMBs, not consumers with predictable patterns and deeply established records. And what happens when fraud strategy design has to account for both uncertainty and growth at the same time.
That is the part I think matters.
Because a strong fintech fraud strategy is not just about tools. It is about understanding the business model, mapping risk to actual customer behavior, choosing the right vendors, and building a fraud prevention roadmap that works in the real world. Matt and Sid get into all of that here, and they do it in a way that is practical, honest, and really useful for teams trying to figure out what good looks like when nothing is especially straightforward.
Here is what that fintech fraud strategy means in practice:
- I need to design controls around the business model, not around generic fraud assumptions
- I need to account for data challenges in fraud when customer information is harder to verify
- I need product mapping for fraud so controls align to real user actions and risks
- I need customer friction reduction without pretending risk will solve itself
What you’ll hear in this episode:
- Why fintech fraud strategy gets more difficult when the customer base does not fit traditional verification models
- How Novo fraud operations approached fraud strategy design for startups and SMBs
- What fraud vendor selection looks like when tools need to match a complex operating environment
- Why product mapping for fraud is essential when different parts of the customer journey carry different risks
- How fraud team lessons from implementation can shape a smarter fraud prevention roadmap
You should listen to this episode if you:
- Work in fraud operations for fintech and need a practical framework for handling complex fraud models
- Are building or refining fintech risk operations for startups, SMBs, or other nontraditional customer segments
- Want better fraud vendor selection criteria tied to operational fit, not just feature lists
- Are dealing with data challenges in fraud and need to make better decisions with imperfect inputs
- Care about startup fraud prevention, SMB fraud risk, and reducing friction without weakening controls
If you liked this episode, be sure to subscribe and review the podcast on iTunes, Spotify, YouTube, or wherever you listen to podcasts. It really helps with getting the word out.
Episode notes & key takeaways
This episode is a really useful look at what happens when a company cannot just borrow a standard fraud playbook and hope it fits. The conversation gets into fintech fraud strategy from the perspective of real operational tradeoffs, especially when customer behavior, available data, and product design create a more complicated risk environment than a lot of vendors or frameworks are built for.
Why fintech fraud strategy has to start with the business model
Let’s break this down.
One of the biggest mistakes fraud teams can make is trying to force a standard playbook onto a business model that clearly does not behave like the standard case. That usually does not end well. And that is really the starting point for this conversation.
Matt and Sid talk about the challenge of building a fintech fraud strategy for a company serving startups and SMBs, where the customer profile is more complex and some of the usual verification signals are either incomplete or just not all that helpful. That changes everything. It changes how I define risk, how I evaluate customer behavior, and how I think about fraud strategy design in the first place.
At first glance, this can sound like a tooling problem. But when you dig in, it is really a business understanding problem first. If I do not understand how my customers operate, what data is missing, and which actions create the most meaningful risk, then even good controls are going to be misapplied.
- Fintech fraud strategy has to be built around the real customer base, not generic assumptions
- Business model fraud complexity changes how teams define risk and expected behavior
- Fraud strategy design works better when product, operations, and fraud teams align early
- Startup fraud prevention and SMB fraud risk require context, not just standard rules
How data challenges in fraud shape decision-making
Here’s what’s actually happening.
A core issue in this episode is that Novo’s customer base can involve unverifiable or less traditional data, which means the team has to make fraud decisions without the clean, familiar signals a lot of other businesses depend on. And that matters. Because when the data is incomplete, the temptation is often to either overreact or underreact.
Neither one is great.
This is where strong fintech risk operations start to look different from simple approval or decline logic. I need to know which signals are actually reliable, which ones are weaker than they appear, and where context can help fill in the gaps. That is a much more disciplined way to approach data challenges in fraud.
Matt and Sid make it clear that the work is not just identifying missing information. It is deciding how to operate intelligently despite it. That is a big difference. A lot of fraud teams spend time complaining about imperfect data. Better teams build around reality.
- Data challenges in fraud force teams to think more carefully about confidence and uncertainty
- Complex fraud models require better interpretation of incomplete or ambiguous signals
- Fintech risk operations improve when teams focus on signal quality, not just signal volume
- Fraud operations for fintech often depend on making better decisions with less certainty
Why product mapping for fraud matters more than teams expect
This is where things get interesting.
One of the most practical ideas in this conversation is product mapping for fraud. And honestly, more teams should spend time here. Because fraud does not show up the same way across every feature, workflow, or customer action. Different parts of the product carry different types of risk, and if I treat them all the same, I am probably wasting controls in one area and missing problems in another.
That is a problem.
Product mapping for fraud helps teams connect actual business actions to actual fraud exposure. Where does onboarding risk show up. Where does payment risk show up. Where are the abuse points. Where is friction most damaging. Where is friction actually helpful. This is the kind of practical work that makes a fintech fraud strategy usable instead of theoretical.
And this is also where fraud tooling decisions get a lot more grounded. Because once I understand the risk map of the product, I can start choosing vendors and controls based on fit, not just on which demo sounded the most polished.
- Product mapping for fraud helps teams place controls where they matter most
- Fraud tooling decisions improve when they are tied to real workflows and risks
- Fraud operations for fintech need controls that reflect different stages of the customer journey
- Customer friction reduction becomes easier when high-friction controls are used more intentionally
What fraud vendor selection looks like in a complex environment
This might not seem like a big deal. But in fraud prevention, it absolutely is.
Fraud vendor selection gets a lot more complicated when the company itself has a more unusual operating model. A vendor can look great in the abstract and still be a poor fit in practice. That is one of the more useful lessons in this episode. The right tool is not just the one with the best claims. It is the one that works with my data reality, my product structure, my engineering constraints, and my operational goals.
Right.
Matt and Sid talk through the practical side of evaluating vendors, and that part is valuable because it moves past generic checklists. Fraud vendor selection should not just be about who has the most features. It should be about who can support the actual fraud prevention roadmap I need to build.
We have seen this playbook before. Teams get drawn toward broad promises, then run into implementation gaps, weak alignment, or products that assume a cleaner environment than the one they actually have. Which, as you might expect, creates a whole new set of problems.
- Fraud vendor selection should be tied to operational fit, not just sales claims
- Fraud tooling decisions matter more when the business model is more complex
- A good fraud prevention roadmap depends on realistic implementation planning
- Novo fraud operations offer a useful example of evaluating tools through a practical lens
Why customer friction reduction is part of good fraud strategy
One thing I appreciated in this conversation is that it does not frame friction as a simple good-or-bad issue. It is more nuanced than that. Some friction is useful. Some is unnecessary. And the difference usually comes down to whether the control is aligned with actual risk.
That is where strong fraud team lessons tend to show up.
If I add too much friction in the wrong places, I frustrate legitimate customers and create drag on the business. If I remove too much friction without understanding the tradeoff, I create openings for fraud. So the real goal is not frictionless fraud prevention in the absolute sense. It is smarter friction. Targeted friction. Friction that makes sense for the risk and the customer.
For fintechs serving SMBs and startups, that balance is especially important. These are customers who often need speed, access, and trust from day one. If the fraud strategy ignores that, the product experience suffers. If the strategy ignores risk, the losses show up somewhere else. Not exactly subtle.
- Customer friction reduction should be based on actual risk, not broad simplification
- Fraud and user experience need to be designed together in fintech environments
- SMB fraud risk requires controls that protect without blocking legitimate activity
- Fraud team lessons often come from learning where friction helps and where it hurts
The big takeaway from this episode is pretty straightforward. A strong fintech fraud strategy is not built by copying someone else’s framework and hoping it translates. It comes from understanding the business model, respecting the limits of the data, mapping risk to the product, and making fraud tooling decisions that actually fit the environment. Matt and Sid do a really good job of talking through that reality. And honestly, that is what makes this conversation useful. It is grounded. It is practical. And it reflects how fraud operations for fintech really work when the business is more complicated than the standard playbook was built for.
