Recent breakthroughs in Artificial intelligence have the potential to revolutionize the software industry. It’s vital that vertical SaaS founders begin thinking critically about the ways AI will impact their productivity, product, customer experience and growth. In the first of a two-part series, we spoke with Andrew Carr, a senior applied research scientist at Gretel AI, an industrial generative AI company focused on synthetic data and privacy. Carr is a member of Tribe AI, a machine learning consultancy and the lead network for top AI Talent, and has worked for Google Brain, OpenAI and his own legal tech startup.

Let’s start with the basics. What is Generative AI?

Generative AI is a term used for Deep Learning models that can generate data. These models are grouped based on the types of data they can generate. The most exciting recent advancements in Generative AI are being made in text and image models. However, there are other models — such as those that generate audio, 3D, or tabular data — which are equally useful, but may have more niche applications.

Within each data type, there are a variety of models to choose from, depending on your organization’s needs and complexity. You can use a hosted model such as ChatGPT from OpenAI, or you can use an open sourced model like FlanT5 or CerebrusGPT. Your choice will determine different levels of customization, cost, and capabilities. These are often referred to as the three C’s of Generative AI.

What are the potential applications of Generative AI in B2B software?

While there are inherent risks anytime you integrate a new technology into an existing product or organization (more on that below) — the potential gains from Generative AI are so exciting that organizations across verticals are moving quickly to adopt language and image generation technology.

Here are some of the most promising areas that we’ve seen companies begin to incorporate deep learning models:

Coding co-pilots — Generative AI and coding co-pilots can improve productivity of software development teams by providing intelligent suggestions for code, identifying bugs, improving code quality, and automating tasks, reducing the overall time and effort for development, and allowing developers to focus on more complex tasks. One Google study showed a 6% decrease in coding iteration time for a team using LLM-powered coding tools, while a Github study showed a 55% faster task completion rate. The wide range suggests a large amount of variability depending on the type of task and uncertainty given the early stage of this technology. But all signs point to a potential for significant future value here.

Customer service and support — Deep learning models can enhance the FAQ and automate customer support, but to avoid generating inappropriate messages, a safer option is to use a language model to summarize support requests into actionable items, which can help prioritize bug fixes and product feature development while improving the customer service experience.

User customization — While not currently feasible, soon generative image and language models will be used to modify the appearance and functionality of a product based on prior user behavior, providing hyper-customized product experiences.

Features discovery — Generative models can suggest new workflows based on a SaaS product’s existing functionality, generating helpful tool-tips or suggestions based on user actions to help them discover new features seamlessly.

Visual question answering — Visual question answering (VQA) is a machine learning technique that uses Deep Learning models to answer questions based on visual input. One example is the Visual ChatGPT tool developed by Microsoft, which combines language and image models to enable real-time answers to natural language questions about images in real time.

Training data generation — Generative AI can help overcome the “cold start” problem by creating a synthetic version of high-quality training data.

FAQ generation — Language models like GPT-3.5-Turbo or GPT-4 can generate related FAQs with answers from company web pages, documentation, and customer service tickets, which can be posted online to improve customer support

What are some of the risks of using generative models and how do I mitigate them?

Generative AI tools are powerful, but come with significant risks that can affect even the most sophisticated organizations. First, generative models can “hallucinate” things (i.e. output data that isn’t factually accurate, but may look convincing.) For example, a well known system was being used to help researchers find papers. The model would make up papers, swap authors, and generally wasn’t able to properly respond to queries. There are ways to mitigate this such as external knowledge bases, but it still isn’t clear that it fully reduces the risk of hallucination.

Second, while the newest models are more aligned than previous versions (meaning the output is more closely aligned with what is expected), they can still produce harmful or irrelevant content. In addition to the initial harm to the user, this can often result in negative PR for the company — a lesser concern, but shouldn’t be ignored.

However, there are ways to mitigate these risks and use generative models responsibly. Current best practices are to have a human in the loop for any AI-generated content. Essentially, a trusted employee signs off on any content before it’s seen by a user.

Another approach is to use AI-generated content only internally within your company. This provides another layer of quality control as the content is seen and filtered by employees before it is used externally with customers or other stakeholders.

A third option is to use specialized models to validate the output of more general models. Because Generative AI models are better at validating statements than they are at generating correct statements, you can use a hierarchy of models to validate the content generated by other models.

What are some of the privacy and legal considerations when using generative models?

When using generative models, it’s important to be aware of privacy and security considerations related to customer data. Regulatory requirements and company processes must be followed to ensure compliance. While many of these considerations are standard across industries and not unique to Generative AI, there are some novel situations that should be considered.

First, some models have restrictions on their use. For example, you cannot use some API models for legal or medical advice. And an open source model’s license might restrict its use to academic research. Additionally, the licensing of pre-trained models must be carefully reviewed to ensure that the original training data was properly licensed for AI training. It’s important to be aware of these considerations to avoid any potential legal or ethical issues.

This week’s essay is by Fractal Operating Partner Joe DiMento

One of Fractal’s core theses is that vertical software businesses can grow much larger than most people thought possible only a few years ago. There are two primary engines driving the expansion of vertical SaaS revenues: (1) businesses in every industry are investing more in software as a percentage of revenue, and (2) vertical SaaS vendors have found new ways to monetize their product with embedded financial services and other offerings.

Even so, the Total Addressable Market (TAM) of many vertical software industries is definitively smaller than those of more horizontal software companies, and this can lead some investors and operators to shy away. But what is often overlooked in this calculus is the relative ease of breaking into a vertical market due to more constrained go-to-market (GTM) options. The vertical GTM motion is much more focused. Teams are able to make much more rapid progress than they could in a horizontal market, and thus have a quicker path to a real, operating company with employees, customers, and revenue.

This may seem counterintuitive, but in fact it’s a classic example of the paradox of choice. When presented with too many options, people are often paralyzed by indecision. They worry about making the wrong choice and spend an inordinate amount of time considering the tradeoffs of a multitude of alternatives. As a result, they lose their ability to act quickly — which is the most important characteristic of a startup founder.

Imagine you’re hungry and you sit down to lunch. You find yourself in a comfortable booth as the wafting smells of delicious meals make your stomach churn with anticipation. A waiter tells you you can have anything you want: sushi, burritos, grilled cheese, souvlaki, spaetzle, mochi — you name it, they have it. Everything is the same price and will be served instantly. What to eat?

Unless you have substantial dietary restrictions or a very narrow palate, this choice will leave you reeling. You haven’t had borscht in awhile, but spaetzle is also really comforting, and has a bit more protein. Then again, you should cut back on fried food, so sushi is a good way to go. But isn’t rice a little bloating on its own? Your waiter is starting to look impatient. If you were standing at a sandwich counter instead, you’d be halfway through your lunch by now.

Horizontal SaaS GTM can feel like this. Except rather than everything having the same price, different paths have wildly different costs. If you want to sell into enterprise customers of highly regulated industries — a bigger moat, harder to break into, but therefore more defensible — your engineering and product team will have to build substantially different features than if you’re going after lightly regulated SMBs. The functional area you sell into — sales, product, engineering, design — will change your new user onboarding requirements. And no matter what Ideal Customer Profile (ICP) you choose, you will always second guess yourself. When your product build takes longer than expected you start to wonder if you should have de-scoped and changed the ICP. You will find yourself being pulled in different directions at times, reactively, based on random user feedback. This totally different type of customer loves the product — should you focus your efforts there instead?

I found this out the hard way in the early journey of selling Coda, the product I worked on for years. We were building a new kind of doc that could solve just about any problem for any knowledge worker, but in the beginning we vacillated on where to focus. We had a small bakery and a large financial services company, both using Coda. Both were great customers! One paid more. The other was a far cooler story. Where to focus?

We eventually got it right by narrowing our focus and diving deep in a few functional areas. But there was a lot of hand-wringing along the way.

Vertical SaaS is the exact opposite. You are given the very simple — and very valuable — gift of focus. This is incredibly important in sales, as our very own Mike Flores outlined in a recent blog post about sales and intentionality.

With vertical SaaS GTM, you don’t need to second-guess yourself along the GTM path because in many ways, that path is already laid out for you.

For example, if you’re selling to scrap yards, like Fractal Company GreenSpark, you don’t have to ask yourself “should we go after municipal waste collectors instead? What about building an app for consumers to track their own recycling habits?” Perhaps these adjacencies can come in the later stages of your journey, but at the beginning you can simply sell to, well, scrap yards. It may not be widely recognized, but this level of GTM focus is a superpower for vertical SaaS vendors and the foundation upon which billion-dollar companies are built.

Not only is choosing where to focus easier — because the choice is made for you — but your GTM channels are so narrowed down as to allow you to avoid the churn of the choice of how to go to market. As an example, when selling horizontal SaaS you can choose from among hundreds of conferences — Dreamforce, SXSW, SaaStr, the list is endless. I vividly remember choosing among a list of dozens of conferences when we were deciding where to visit at Coda. There’s that food menu analogy again. When selling to scrap yards, the CEO of GreenSpark chose the Institute of Scrap Recycling Industries conference, or ISRI for those in the know. “Conferences are our largest revenue driver,” he said, of the channel. Unlike horizontal SaaS, there aren’t hundreds of scrap metal recycling conferences to choose from.

Designing a landing page? Well, who’s your actual buyer? If there is one audience, it’s easy to make sure what they care about is mentioned on the page. You don’t worry about anyone else! Prospecting to sales targets? I spoke to a Fractal CEO the other day who is selling into local government. Their lead building strategy is simple: comb through all the public websites and find the names, emails, and addresses of the functional role they’re selling into. There are hundreds, if not thousands of leads, out there on the internet, waiting to be found. And crucially, the CEO will never have to second-guess where to look for them.

In this way, I often describe GTM for vertical SaaS as “not easy, but simple.” It takes a lot of work, grit, elbow-grease, name your analogy here — but there is very little of the mind-bending complexity of layering on ICP with geo with function with industry mapping that a horizontal SaaS company has to do.

To return to a version of my first analogy: Instead of becoming the head chef of a Michelin star restaurant and having to choose the cuisine, location, interior design, and of course, menu of a new restaurant venture, Fractal CEOs begin their GTM journey as line cooks. They dutifully mince garlic and julienne carrots before their chef careers take off and have the opportunity for more complex efforts.

Keep doing the important but unglamorous work of early-stage SaaS sales — cold calling, pitching, and closing deals. When you pick your head up in ~9 months, you’ll have a growing company with real customers, real revenue, and real results.

That’s the gift of focus. It enables vertical SaaS companies to achieve unbelievably rapid progress. I’ve spoken with founders in our portfolio who shook hands with their cofounder on Day 0, and on Day 130 are collecting revenue from customers for a product they have already shipped. Now that’s fast.

This week’s essay is by Fractal Senior Associate Gabriella Monico.

Over the past couple of weeks, I had the opportunity to chat with several Fractal founders building at the intersection of vertical SaaS and healthcare. All of these founders have gone through key early milestones, including launching their MVP, onboarding early customers, and raising institutional capital. While they are building in distinct verticals, I found significant overlap in their experiences.

We discussed scoping an MVP, building credibility with clinicians, and other common challenges healthcare SaaS entrepreneurs face. Here are some of the most salient takeaways from each of those conversations.

1. Switching costs in healthcare are high, so a new player must be significantly differentiated against competitors to drive adoption

When CEO Braedan Russell started Exacare, an electronic medical record (EMR) and practice management software for assisted living facilities, he knew that EMR systems are notoriously sticky. While users may find their existing solution subpar, challenges associated with data migration and user onboarding often prevent them from replacing existing solutions. This makes it difficult to garner significant adoption by offering marginal improvements like better UI/UX or more effective reporting capabilities.

To convince users to switch to a new provider, it has to significantly improve their way of working. “Your product can’t be 2x better,” said Russell. “It has to be 5x or even 10x better than the incumbent.” One of the key differentiators for Exacare was saving nurse’s time by drastically improving resident documentation capabilities. The nursing shortage made this feature particularly compelling to assisted living facilities.

Walker Szurek, CEO of Prospyr, which builds software for aesthetics practices, agreed with Russell’s insight. According to Szurek, Prospyr created a competitive advantage by allowing users to share images to social media directly from the EMR. “You have to find the silver bullet that you don’t see anyone else do well,” said Szurek. With Prospyr, aesthetics practices were able to bridge the gap between their EMR and marketing software, a value-add specific to plastic surgeons and medical spa operators that no other EMR provided.

2. The most effective method for finding a wedge is spending time at practices observing clinicians at work

Finding a wedge or feature set that will ultimately convince customers to switch to your product is not an easy task for founders, especially in a healthcare industry where outdated legacy practice management incumbents offer a host of features. It can take months of iteration to identify what really needs to be included in an MVP, and Szurek says that Propsyr determined the most valuable features by spending multiple days inside practices, active listening, and observing clinicians engaging with their software.

Noah Whitehead, CEO of Ritten, an EMR for behavioral health clinics, said he “spent months at the facilities of early adopters.” All that time was crucial to scoping his MVP and led to Ritten deprioritizing some lower-value features that he once believed were crucial, such as a complex reporting feature.

By observing clinicians at work, engaging in hundreds of hours of conversations with customers and advisors, and relying on a keen sense for product development, Szurek and Whitehead were able to empower their respective teams to build a differentiated MVP with a significant edge over their competitors.

3. Leverage well-regarded early adopters to build credibility with future customers, especially if you lack a clinical background

Like the majority of the healthcare founders in Fractal’s portfolio, Whitehead did not have a clinical background. Consequently, his team relied heavily on feedback from early customers and advisors to build their product. In doing so, they developed strong relationships with those facilities, and this proved crucial in building credibility with future customers, signing new logos, and improving sales efficiency.

“What really helped is that we have these two real licensed brand-name facilities that signed on very early,” said Whitehead. “When they could provide a reference for others, we pointed everyone to those facilities. That was the biggest game changer.”

Investing in the success of early customers may increase CAC in the short term. But as Ritten’s experience shows, it can pay huge dividends down the road when these customers become outspoken champions of your product.

Mike Flores leads sales strategy for Fractal’s portfolio of vertical SaaS companies. His guidance enables founders to secure early adopters, build a repeatable sales motion, and deploy a pioneering sales team. In this essay, featured in our recently published Guide to Growth, Mike provides insights and lessons for founders looking to scale their sales organization.

Over the past decade, I’ve watched a pernicious idea take root in the startup world. It’s an idea that suggests that scaling a software company is all about finding shortcuts. That if a founder can pull a broad list of emails and press send on an eight-step outbound email sequence, they can conjure some initial sales traction. It reduces the hard work of a high-functioning sales org into an exercise in experimenting with cheat codes, as if mashing buttons in the right order is the key to attracting and retaining customers. The truth is there are no shortcuts in sales.

If my experience with scaling B2B SaaS sales teams has taught me anything, it has taught me that the difference between a successful high growth startup and a floundering startup often comes down to a single factor: intentionality. They are intentional about the customers they pursue and how they pursue them. They are intentional about the sales reps they hire and the way they structure their team. They are intentional about the tools they use and the data they collect. Startups that successfully scale give the finger to fate and proceed according to their plan. They are intentional in everything they do.

From Early Adopters to a Scalable Sales Motion

Once a CEO has developed a perspective on their industry, validated it by closing early deals, and launched their product with these customers, they’re ready to start scaling their sales organization. This inflection point will be different for every company, but eventually the CEO will become the bottleneck in the deal pipeline. The CEO’s task is now to find a few account executives that will lead the way on scaling their sales motion, starting with early adopters. These people are not easy to find. In fact, the CEO and early account executives should become experts on disqualifying opportunities. This seems counterintuitive, as the sales pipeline is hard enough to build. However, in my experience, you have to find early adopters who will persevere through the inevitable rough patches, and will put your project at the top of their priority list.

If the CEO and their first sales hires have successfully threaded the needle and secured a healthy cohort of early adopters, the exercise shifts to pattern recognition. This is by far the most important skill for enabling fast growth. A repeatable sales motion relies entirely on narrowing down an ideal customer profile that represents a broad enough selling opportunity, and then using those customers’ own words to attract more similar companies. Rather than spraying lead lists and hoping something sticks, scalable sales requires the outbound team to unearth non-obvious triggers to find comparable customers. This is the art of intentional prospecting or account based sales.

Intentional prospecting is a trend we’ll see gain even more traction in years to come. At minimum, a cold outbound email has to demonstrate at least two things to the recipient:

1. why they (and only they) could have been the recipient
2. why you have the conviction that they are at (or near) an inflection point that would require adopting your solution

Doing this at scale is hard. It’s why so many software sales organizations opt for volume-based sales strategies rather than putting in the effort that account-based sales requires. However, evidence from organizations I’ve worked with show that the win rate of an opportunity sourced using an intentional prospecting methodology is 3x higher than an opportunity sourced via an antiquated, cadence-based mass email method. Founders are duped into playing the volume game because it often seems to be working at first. Sales reps will land customers and meet quotas, but what they haven’t done is build a scalable sales process. They are closing customers essentially at random, which isn’t the foundation for sustainable growth. The only way a founder can tell that they have a scalable sales process is by being very intentional about the metrics they track.

Measure Everything

The trouble with playing the volume game in sales is that it doesn’t result from intentionally targeting the right customers with the right product at the right time. There are no learnings here that can be used to further increase sales because account executives don’t really know why they are closing prospects. Answering this question requires diligently tracking fundamental sales metrics like outbound conversion ratios and account penetration; I am constantly amazed at how few startups track even these basic sales metrics.

Without this data it is impossible for the CEO to make informed decisions about the company’s sales process. This leads to a common and fatal error at many startups: rapidly scaling their sales org before they’ve established a repeatable and predictable sales process. Without the numbers to understand whether a sales process is working, it’s easy for founders to believe that simply hiring more sales reps will result in more sales. This is a volume-based sales mindset. But if the current sales executives have low conversion rates or if there is unsustainable customer churn, this means the sales process is fundamentally broken. If the CEO hires more account executives before refining their process, this will only compound the problem and lead to failure faster.

Being intentional about the sales data that is being collected and the metrics that are being monitored prevents the startup from relying on guesswork to grow. Instead, a company culture based on intentionality and metrics allows the CEO and sales leaders to precisely diagnose and fix the problem. It also ensures that the solutions themselves are both sustainable and effective. For example, if a startup has a monthly review of conversion metrics and finds them to be too low, the temptation is to sometimes look at new tooling or methodologies to correct the course. I’ve certainly been guilty of this, especially when our budgets seemed infinite. But the days of growth at any cost are long gone. Today, only companies with provable sales efficiency are going to survive.

As a founder, taking a beat to come to a consensus about the root problem is an important first step. Sometimes, new tooling is indeed the solution. But more likely the solution will require more clarity around the problem you’re trying to solve and how the buyer should embark on a journey with you to solve that problem. Once you understand this, it is relatively straightforward to identify ideal customers and create a scalable account-based sales motion. With all the data and tooling that is available on the internet, there’s really no excuse to not do intentional prospecting. The companies that don’t put in the effort are only admitting that they don’t really know who their customer is and the value they provide to them. The companies that do will understand their customers at a much deeper level and will build a scalable, high-efficiency sales machine.

Cultivating intentionality

The challenges with scaling a sales organization don’t stop once a company has proved to itself that it has created a repeatable sales motion. As the company grows, its ideal customer profile will inevitably change and it must adapt its strategies to the needs of the new customer profile. At the same time, the CEO must contend with the hiccups that usually accompany passing the reins to a new sales management strata within the organization. They will be forced to make difficult decisions about how and whom they hire. Should they promote an internal AE to head of sales and risk losing one of their best reps, or should they risk taking on an outside hire who will be coming into the company blind. There are no one-size-fits-all solutions to these types of management challenges, but I can guarantee you that it will be easier to solve them if the CEO has created a culture of intentionality in terms of how AEs source prospects and track metrics.

My belief that a startup should approach everything it does with intention may seem painfully obvious. But all too often startups proceed without much of a plan. This is typically because they don’t deeply understand their industry or customers, and so they are willing to try anything and everything to boost sales in the hope that they will eventually stumble upon a solution. As a startup grows, so does the complexity it must contend with. Addressing this complexity with intention becomes more challenging and founders will be tempted to avoid the near term pain of creating a plan and opt for ad hoc solutions instead. This may work for a while, but in the long term it will stunt the growth of the company and lead to even greater pains down the road.

Bobby Tzekin is CEO & Co-founder of Wisetack. Founded in 2018, Wisetack is the leading Pay Over Time platform for in-person services. Wisetack partners with top SaaS service companies to embed payment options directly into their software through a simple API integration. Small service businesses rely on Wisetack to offer their customers flexible, point of sale loans, enabling them to quickly and easily pay over time for essential services.

Most people wouldn’t think of “consumer lending” as relevant in a B2B context. Why is this kind of embedded financial tool valuable for vertical SaaS companies?

For the vertical SaaS companies we work with, consumer lending can bring a lot of value. The reason for that is that these companies have two audiences: the businesses who pay to use their product (let’s call them “merchants”) but also those businesses’ customers (“consumers” or “end users”). For example, in the home services vertical, the merchant could be an HVAC or landscaping company and the end user would be a homeowner.

When there is an end user, creating a positive experience for them (e.g., simplifying how they book service with the merchant) benefits all parties: the end user is happier, which is good for the business that serves them, which is then good for the vertical SaaS company that serves the business. We like to call that a win-win-win, and we believe consumer lending can accomplish this. By helping the consumer pay over time for merchant services, you help the merchant get more sales and more loyal customers. And, along the way, you can help merchants even more by doing things like ensuring they get paid promptly and providing resources to help them get the most out of their financing program.

What are some industry characteristics where it makes sense for a B2B vertical SaaS company to offer consumer facing lending options to their customers? Is this more effective for some industries than others?

Average transaction size within the merchant base is probably the most obvious characteristic to look at when a B2B vertical SaaS company is considering offering a consumer-facing lending option. For consumers, financing is more attractive for amounts that fall outside of their normal expenses, so merchants offering such services might benefit the most from offering financing. That said, there are many, many situations where the consumer chooses to finance with us primarily for convenience, and not just to help smooth out their budget, and those use cases can happen regardless of transaction size.

At what point in a vertical SaaS company’s growth should they start to think about servicing their customers’ customers with embedded consumer lending? What value are they driving to their customers with this kind of tool?

It’s crucial to start thinking about embedded financial products early on. That being said, this area might not be the very first priority for a vertical SaaS company. It takes a certain level of trust for customers to adopt a financial product, and for a really early stage company, building this trust may take some time.

Generally, a good starting point is to add a card processing option, since there are so many great products that are relatively easy to integrate. Once people trust you as the platform through which they get paid, the door is open to many other products — and embedding consumer financing is a great next step.

However, there are scenarios where it makes sense to think about financial products before card processing, like when a SaaS is focused on the lead gen/marketing/sales interaction between merchant and consumer. If they are driving marketing and sales, then financing arguably becomes more relevant than cards (e.g., Thumbtack and Yelp).

From our experience, the most successful embedded fintech/lending strategies come from companies that have a person (or people) dedicated to fintech, which implies a certain level of commitment to an embedded fintech strategy. Most early stage companies won’t have this in place, but for growing companies, creating this role is something to consider.

What common challenges or misconceptions have you noticed about embedding consumer lending products in B2B software?

Many vertical SaaS companies miscalculate the level and timing of investment needed to make an embedded fintech product successful. On one hand, it can be relatively easy to get an initial implementation out the door — easier than many SaaS companies might think. On the other hand, following the implementation, a substantial investment is required to maximize the potential of the offering. Those who don’t put effort into sales, marketing and customer success fail to drive adoption are often disappointed in the results. Companies also need to anticipate the investment required to build future iterations of the embedded product. This will be driven by customer feedback, and it’s a key part of driving adoption.

How does embedding consumer lending capabilities into a product offering for B2B vertical SaaS companies actually work? What does it look like in terms of implementation, partners involved, risk borne by the vertical SaaS company, risk borne by their customers, etc.?

Providing a universal answer to this question is challenging, as there are various models for this. But in the case of Wisetack, we try to make it as simple as possible to get up and running in terms of implementation, user experience, and risk.

Typically, on the partner end, the project involves one product owner (usually a product manager or fintech business lead), and one or two engineers. In terms of the timeline, we have seen integrations completed in as little as three weeks — including all design work, engineering, and final testing.

Our APIs are easy for most SaaS companies to integrate into their products, and flexible enough to fit into an existing product without significantly altering UI/UX and core workflows. One of our philosophies is to meet partners where they are. For example, when it comes to designing the UI, we can provide as much or as little support as needed. We can propose entire UI designs, but if the partner has a large design team and wants to own it themselves, we are happy to just step in where needed — or not at all.

We apply the same principle when it comes to risk. For most partners, we fully own the risk (for both the partner and merchant). But if the partner wants to share in the risk to share in the upside, that is something we can support too. In either case, this removes the need for SaaS companies to build up their internal risk functions to launch lending, and makes it easier and faster to get the program up and running in a scalable way.

We just released our Guide to Growth, which explores the challenges and opportunities software companies and their founders face as they prepare to scale their business. In a series of essays written by former founders and senior operators, we look across product, recruiting, sales, and engineering to better understand how to succeed when it’s time to accelerate towards $100M of ARR.

For far too long, vertical SaaS companies have been unfairly characterized as businesses with limited growth prospects. By focusing only on first-order addressable markets, both founders and investors couldn’t see the many opportunities for vertical software companies to dramatically expand their TAM by layering in fintech products, marketplaces, and other potent revenue drivers. They ignored the unique leverage enjoyed by vertical SaaS companies that comes from controlling the key workflow processes of their customers and the potential for these companies to own the majority of software spend for entire industries. At best, a venture-scale vertical SaaS company was the rarest of beasts; at worst, it was nothing more than a myth.

At Fractal, we’ve always believed that vertical SaaS companies have the potential to become large, multi-billion dollar businesses if they have the right guidance and support. We don’t help launch companies in the hopes of their founders hitting a single or a double. We believe that every vertical SaaS company in our portfolio has the opportunity to hit a homerun. The runaway success of companies like Toast, Epic, Procore, ServiceTitan, Veeva, Qualia, and many others have revealed that there are untold numbers of (multi-)billion dollar vertical SaaS businesses waiting to be built.

Scaling a vertical SaaS business requires founders to successfully navigate a treacherous growth phase that is filled with unexpected challenges. Growing a company isn’t simply a matter of doing more of whatever you were doing before. The success of a startup during its growth phase largely depends on whether its founders can create well-defined and repeatable — in a word, scalable — processes at every level of their organization.

That doesn’t mean these processes must be rigid. As a startup scales the processes that drive the organization can — and should — adapt to the new demands placed on the company. For many founders, implementing new processes can be a painful experience. After years of hacking and improvising to land customers and ship code with a small team of early employees, even the most flexible processes may feel rigid and superfluous. But as a company starts to grow, these processes become critical to maintaining forward velocity. Without them, internal communication breaks down, product quality declines, and customers churn.

Fractal’s leadership has deep experience navigating the unique challenges and opportunities that arise during the growth phase of a company that they use to inform the guidance they provide to founders in our program. In this guide, we’ve distilled some of their wisdom into 4 essays that explore the challenges of growing a post-Series A SaaS company, including:

• how a CTO’s role changes as their engineering team grows
• the careful balancing act that comes with hiring your first product manager
• how to build a sales team with intention
• how to hire the best employees for a job you’ve never done

The insights from each essay are derived from years of experience driving growth at companies like Lyft, Snapdocs, Scout RFP, Ironclad, Workday, Namely, and Indio. They contain advice that Fractal’s leaders wish they would have known when they first experienced growth pains at their own companies. We hope that founders everywhere find them useful on their journey to building the next generation of massive vertical SaaS businesses.

READ THE FULL GUIDE TO GROWTH HERE

This week’s essay is by Fractal Senior Associate Gabriella Monico.

The conventional wisdom for vertical SaaS investors and founders is that a successful industry software company should aim to serve most — if not all — of the unique workflow needs of their core customer base. That’s typically the reason investors are comfortable backing founders building companies with smaller first-order TAMs — there is opportunity to extend the product up the value chain and grow ARR with minimal sales effort. The roadmap varies, but vertical SaaS companies often start with a wedge product or point solution that they can grow into a full system of record. After building a system of record, the company may then develop a vertical-specific system of engagement or a fintech offering or a marketplace.

It’s a tried and true playbook followed by successful vertical SaaS businesses like Squire and MindBody that service the barbershop and boutique fitness industries, respectively. But some verticals are just too complex to be effectively serviced by a single vertical SaaS provider. Internally, these businesses often need to purchase and integrate multiple software systems to effectively manage their diverse workflows. Externally, they find themselves enmeshed in a network of service providers, brokers, and customers that all use different software platforms.

As businesses digitize and grow their tech stack with vertical and horizontal software solutions, they quickly realize the importance of interoperability across platforms. Although some verticals are too early in their digitization journey or lack the operational complexity required to merit multiple vertical and horizontal solutions, there are many industries that already greatly benefit from vertical-specific API solutions (e.g., construction and finance) and demonstrate their powerful enabling effects.

Generally speaking, industries that may benefit from verticalized API solutions typically share the following characteristics:

• Strong digitization tailwinds
• Multiple mission critical, but siloed software systems
• Entrenched in a complex network of stakeholders that require communication and data sharing for successful operations
• Strict regulations around data sharing

Two particularly notable examples of industries that have all these characteristics are healthcare and logistics, which have both recently launched discussions about software interoperability to the forefront of national discourse.

A Case Study in Healthcare

Multiple regulatory bodies in healthcare including the Department of Health and Human Services (HHS), the Center for Medicare and Medicaid Services (CMS), and Office of the National Coordinator for Health Information Technology (ONC) have worked in tandem to establish standards for healthcare interoperability. The ONC Cures Act Final Rule promotes patient control over health information through standards-based API adoption. In addition, interoperability is particularly important for hospitals receiving incentive payments through the CMS Promoting Interoperability (PI) program.

Healthcare API companies like Redox Health and Lyniate were built in response to the growing importance of interoperability in healthcare. These companies partner with providers to connect multiple systems including EHRs, clearinghouses, billing vendors, remote patient monitoring tools, and more. This facilitates financial workflows, enables the transition of care, and improves remote patient outcomes. In addition, payers can leverage these tools to access EHR and patient intake data across providers in order to improve rates of prior authorization, identify gaps in care, and retrieve medical records. They also empower software companies that service healthcare organizations because they can partner with these APIs to access clinical and financial data across platforms.

As interoperability standards improve and value based care adoption accelerates, APIs like Redox and Lyniate will be a necessary tool for all health systems to be in compliance with national standards and increase profitability in value-based reimbursements.

APIs Revolutionize the Supply Chain

Healthcare is not the only industry waking up to the importance of software interoperability. . The Covid-19 pandemic unveiled significant challenges present in supply chain management and resulted in shortages of essential goods, price inflation and factory closures.

In a survey of 200 senior supply chain executives, supply chain visibility was their top business priority in 2022. Unsurprisingly, this demand has led to significant attention and investment in vertical specific APIs for supply chain visibility like Project44 and Vizion. These solutions push live container tracking data to ERPs, TMS systems and other solutions leveraged by a shipper or logistics business.

Like Redox Health and Lyniate, these supply chain visibility APIs enable mission critical workflows including efficient inventory allocation, improved on-time delivery, and data-driven port selection. By allowing shippers and logistics providers to track container location in real time, supply chain APIs are poised to help solve the most vexing problems in logistics. And as vertical SaaS businesses continue to scale from the accelerating digitization of their industries, we expect to see a strong class of new vertical API vendors ready to solve their interoperability challenges.

Mike Barbosa is CEO and Co-Founder of OatFi. OatFi provides the end-to-end infrastructure for B2B Payment and Vertical SaaS platforms to embed and monetize working capital tools for commercial end users. Through an integration with OatFi, SMBs are empowered to collapse their cash flow conversion cycle, and platforms receive a new, high-margin revenue channel at zero-added risk

FRACTAL: Tell us about the embedded lending ecosystem. What sorts of technology solutions are involved and where does Oatfi fit into that ecosystem?

Mike: The embedded lending ecosystem allows software companies that aren’t traditionally lenders to offer and monetize credit products. Across the ecosystem, you can break the main players into two categories, with some crossover depending on product strategy: infrastructure and embedded lenders.

Infrastructure companies provide tools that enable SaaS companies to build their own products. These companies serve functionality across the credit stack, including Know Your Business (KYB) , underwriting and risk decisioning, ledgering and loan ops, capital markets, collections and servicing, and risk management. While each piece of infrastructure serves its own niche, there is a considerable operational lift to manage the workflow owned by each category of vendor.

Embedded lenders are companies that provide a holistic product offering for SaaS companies to deploy credit. We can bifurcate this category into two groups: those that manage the UX/UI and those that don’t. Within the first category, some require their brand to be seen, and some do not (think Affirm vs. Lendflow). For the embedded lenders that don’t provide the UX/UI, SaaS companies are empowered to build their own offering on top of the lender’s credit stack.

What problems are embedded working capital tools solving? Please give an example (hypothetical or real) of how this kind of infrastructure helps a SaaS company.

The most critical problem for businesses is cash flow. The unfortunate reality is you can have an amazing gross margin, but if you don’t have cash available you’re going to have a very hard time running and growing your business. Oftentimes, the cashflow problem is not uniform across industries and is dependent on supplier or buyer-specific payment terms.

Offering both embedded accounts payable and accounts receivable solutions empowers businesses to solve their cash-flow-conversion cycle no matter where the issue lies. — Embedding working capital tools into SaaS offerings enables companies to choose how and when they want to pay and get paid. It eliminates mismatched terms. Here are two examples of how that works:

Accounts Payable Example: Assume you are a SaaS solution serving the needs of the security guard-staffing industry. The biggest issue for many of your users is payroll. While payroll at most companies is biweekly, this cadence can be difficult when customer invoices are paid 30–60 days after service completion. This scenario helps us understand a common mismatching of payment terms (i.e., when cash comes in the door vs. out the door). If this staffing firm is able to access an AP financing tool, they can align their payroll expenditures with their incoming revenues. In this case, the staffing agency is able to solve their working capital issue by extending their days to pay, otherwise called their ‘Days Payable Outstanding.’ For most businesses, the higher this number, the better!

Accounts Receivable Example: Looking at the same staffing company, we can view the mismatched payment terms from their accounts receivable where they have to wait 30–90 days to get paid. The number of days they wait for outstanding receivables to turn into available cash is called ‘Days Sales Outstanding.’ Businesses typically want this number to be as low as possible. These receivables are typically paid for jobs and services already completed, but since industry-wide payment terms are net 30–60 days, staffing companies are forced to cover operating expenses from their cash on hand. This is where an AR financing tool can be offered. Embedding a factoring-like product into this SaaS solution allows staffing companies to receive funding upon invoice creation and access the cash needed to fulfill their biweekly payroll and operating expenses.

In terms of the kinds of SMBs that our vertical SaaS companies are serving, are there certain types of customers or industries that particularly benefit from this kind of solution? What are their characteristics?

Most SMBs face cash flow issues, which means the need for working capital tools industry agnostic. Whether it’s because of industry standard net-terms, build time, or project length, cash flow issues can affect businesses of all shapes and sizes.

How do embedded working capital tools enable loans without requiring SaaS companies to create a credit product? How does this work on the backend? What is the integration load for a SaaS company’s engineers?

Embedded working capital tools aim to package the intricacies of a credit product into API and SDK integrations. Depending on the lending platform’s APIs, these integrations can be completely white-labeled to the SaaS platform or alternatively branded as the lending partner, introducing an additional workflow. For example, here at OatFi, we enable SaaS platforms to build their own product offering from a UX and branding perspective, while we own the lift of originating the loan both from a technical and capital markets perspective. In our case, OatFi is the lender of record. So when a SaaS platform’s user takes a loan, they are not borrowing from that SaaS company, they are borrowing from OatFi.

On the backend, most embedded lenders originate a loan by receiving data for underwriting through a hosted onboarding flow surfaced by an i-frame, while other platforms — includingOatFi — leverage APIs to retrieve this information. This data is then able to fuel near-instant underwriting decisions, which are then sent back to the end-customer through the SaaS platform’s interface. Finally, once lending terms are agreed upon, funds are either pushed out directly from the lending partner or pushed through the partner platform onto the end-customer.

What should a vertical SaaS founder consider before they decide if this kind of tool can help their business?

Vertical SaaS founders know their respective industries inside and out, so no one is better equipped to understand the nuances of their customers’ cash flows. If a lot of discovery calls with your customers revolve around the time-lag when making payments or getting paid, embedded lending needs to be on your roadmap. Embedding both payment and lending functionalities at once helps founding teams differentiate themselves from the market and hit the ground running with a full suite of fintech offerings.

Why is embedded working capital a high leverage tool for vertical SaaS companies in particular?

Working capital tools like OatFi give vertical SaaS companies a 100% gross margin revenue stream, while empowering your SMB customers to operate more efficiently. By embedding payments and lending products directly into your platform, SaaS companies can increase user retention as they capture more of their customer’s workflow.

Investors often refer to vertical SaaS vendors as “niche” software companies, which can mean different things depending on their opinion about industry-specific software. Vertical SaaS bears will interpret niche to mean that these software businesses have a small total addressable market and limited upside potential. Vertical SaaS bulls see niche as an important competitive advantage that seeds the growth of massive software companies. The bear’s perspective is understandable, but that doesn’t make it less wrong.

For decades, companies pursuing large TAMs have been the darlings of venture investors because of the size of their revenue opportunity. As Sequoia Capital founder Don Valentine once put it, “I like opportunities that are addressing markets so big that even the management team can’t get in its way.” It’s been the guiding principle for VC investing ever since.

Valentine’s focus on large TAMs launched the tech companies that made Silicon Valley into what it is today. But as the software industry matures, over-indexing on first-order TAM increasingly leads to suboptimal investment decisions. What this perspective downplays is that larger TAMs mean a lot of competition. In these markets, winner-take-all dynamics rarely emerge and if they do they frequently attract antitrust suits based on the sheer scale and power of their monopoly.

The result is typically vicious and unending competition just to own a fraction of the market that is constantly at risk of being lost. SaaS companies competing in these markets must spend small fortunes on sales and marketing, which can quickly demolish their otherwise exceptional gross margins. This forced S&M spending kills their operating margin, which ultimately kills their profitability. If the long term goal of the company is to be acquired or go public, free cash flow is what really matters.

Vertical SaaS companies, by contrast, operate in “niche” markets that are typically defined by industry workflow needs. While the first order TAM of these SaaS businesses is relatively small compared to, say, a CRM company like Salesforce, this vertical-specific focus is advantageous because these software companies have both fewer competitors and more focused needs, enabling a higher velocity early go-to-market operations. Moreover, vertical markets do tend to result in winner-take-all dynamics which means that a company that develops a vertical SaaS product can create a de facto monopoly in their vertical.

High market share comes with a lot of benefits to the company, including pricing power and lower customer acquisition costs. This results in significantly lower operating margins and more free cash flow, which the company can use to launch expansions into other markets or complementary products that can increase its TAM by many multiples. The leverage that vertical SaaS companies can create by operating in niche markets is the key to turning these seemingly small markets into multi-billion dollar opportunities.

The past ten years have shown us that time after time first order TAM has proven to be a misleading indicator for the size of a business’ opportunity. If investors had just looked at the first order TAMs for companies like CoinBase, Uber, or WhatsApp, they would’ve missed out on the incredible growth opportunities for companies competing in what appeared to be laughably small markets. The same is true for vertical SaaS. Companies like ServiceTitan, Toast, Procore, and the swelling ranks of vertical SaaS unicorns show that going niche is a shrewd pathway to building software businesses that are anything but small.

In late December, hundreds of thousands of people found themselves stranded at airports around the country due to a meltdown at Southwest Airlines. Between December 21 and December 31, the budget airline canceled nearly 17,000 flights due to a “once in a generation” winter storm that impacted nearly every state in the US. The debacle is expected to cost Southwest nearly $800 million, but this massive figure still doesn’t really capture the immense hardships borne by the company’s customers. These were people who had their lives upended for more than a week, missing vacations, work, and precious time with their families that they’ll never be able to recoup.

Although last month’s winter storm caused significant cancellations at every major US airline, all of these companies promptly bounced back after a few days — except Southwest. So why did one company descend into chaos while its competitors took to the skies? The answer is inadequate software, a fact that dramatically underscores the importance of vertical SaaS in our daily lives.

Southwest relies on customized software to handle flight scheduling, but what few people outside the company realized until their flight was canceled over the holidays was just how outdated and fragile that software platform really was. As the president of the Southwest Airlines Pilot Association told the New York Times, Southwest was using “I.T. and infrastructure from the 1990s” that was overwhelmed with the volume of canceled flights. The software lacked even basic functionality such as a simple automated way to contact and reassign pilots and crew members to new flights, which is a standard feature of software used by other airlines.

It’s difficult to believe, but in 2022 the only way that Southwest could reassign most of its crew members was by individually calling them or having someone at an airport track them down to tell them. This is a problem that modern workflow software could have easily solved.

Southwest’s software challenges have been known for years, but have largely gone unaddressed. In 2021, the company suffered two technology meltdowns that resulted in thousands of canceled flights. Even the company’s CEO Bob Jordan acknowledged that their software was inadequate for the demands of a modern airline. “We’re behind,” he told Fortune just a few weeks before Southwest’s recent fiasco. “As we’ve grown, we’ve outrun our tools.”

If there’s any lesson to be learned from Southwest’s meltdown, it’s that modern vertical SaaS is critical to the health of our nation’s economy and many of the day-to-day comforts we take for granted. It’s natural to assume that companies that provide critical transportation infrastructure would have access to technologies that enable them to effectively deliver their services. Southwest is a $20 billion company, after all. Surely it can afford some decent software?

Yet what Southwest revealed in a particularly dramatic way is just how common it is for businesses to use outdated software to deliver essential services. While this is true even of multi-billion dollar companies like Southwest, it is particularly true of the tens of thousands of small businesses that are the bedrock of our communities and the US economy. These local businesses are just as essential to our daily lives as major corporations like Southwest, and they deserve software that will enable them to provide the services their customers depend on. Southwest could afford to build its own custom software in-house, but this luxury isn’t available to most small and medium-sized businesses. In many industries, a high-quality SaaS platform tailored to these businesses’ needs doesn’t even exist.

This is a problem we set out to solve when we founded Fractal. Our goal has always been to empower the next generation of software entrepreneurs to build modern tools that serve the needs of every industry in America. We believe that vertical SaaS isn’t just a great business model — it’s a necessary one. As the thousands of people who couldn’t make it home to their families over Christmas learned the hard way, the types of software businesses use can make a material impact on all of our lives. We want to make sure that these businesses are only using the best.