Discover more from Scatter Brain
My chat with Thomas Brunskill, Cofounder & CEO of Forage
Transitioning from hire-then-train to train-then-hire model of employment
This is the most comprehensive chat I have had so far. I talked to Thomas Brunskill, the CEO of Forage, a company I have been following for a few years. Forage was founded in Australia with the mission to transform how employers hire people and help candidates explore career paths before committing to jobs. I remember thinking they had a clever model when I first discovered them. I finally got to explore this multi-faceted business that deserves an investigative breakdown. Towards the end, Tom shares a thought-provoking perspective on being an immigrant founder in this new environment, what he wants his legacy to be, and what he believes we are not paying enough attention to.
Sar: Helping people find a job and learn new skills to find employment on the internet has been a massive market for the past 20 years. The first generation of companies tackled those two problems separately and typically took a marketplace approach. There are marketplaces of job listings that help you find and apply for jobs, and there are marketplaces of (user-generated) courses that help you learn new things. In both cases, the companies took a lightweight approach of making the discovery and matching for both sides and took on other aspects like course payments and job applications. The heavy lifting of finding the right jobs and courses that align with one’s job search criteria or learning goals was left to the users.
The second generation of companies started combining the two problems. They would help you learn the right things by taking on curriculum design and delivery while also helping you get placed in jobs. These companies started verticalizing by tackling a skill like sales or coding to customize their offerings. They trade schools on the internet.
Before we get into what you guys do, do you agree with my read of how the market has evolved? I think the transition between the generations highlights the challenge and opportunity of matching people and jobs better. Companies are taking on more and more of the work to generate better matches while trying to solve the disconnect between job outcomes and business models of the traditional university system.
Tom: Yes - I agree with that read, although there are a few important threads to pull out. Effective job matching is a very hard problem - many have tried and failed. But we are getting closer to the answer with each generation of companies.
The first generation of companies opened up access to opportunities through digitizing the job application process and skill-building experiences. But in some respects, those early job listing websites worsened the job matching process. The early job-listing giants (e.g., Monster, Indeed, Seek) applied popular product principles to the job application process. They focused on removing friction from the job application process, which led to the ‘one click’ application phenomenon. But friction in the job application process is a good thing --it’s needed. Recruiters don’t want to sift through thousands of applications with limited signals on an applicant’s intent and capability - a problem that those early companies only exacerbated. On the edtech side, the first gen companies (e.g., Coursera, Udacity, etc.) opened up access to knowledge that historically was difficult to access (because you had to go through a four-year college degree). There was this extreme proliferation of educational content, albeit with significant variability in quality. But these companies missed that knowledge alone doesn’t get you a job. An employer’s trust in a credential gets you a job, not the knowledge you gain from a course.
I characterize the second generation as companies focused on establishing their credential with the explicit goal of that credential leading to an employment outcome. This approach was an improvement from 1st gen companies, but they struggled to get credible buy-in from employers and candidates. These companies took on the monumental challenge of trying to convince employers to hire using a new credential from a new and unknown provider. It’s (almost) impossible to get an employer to adopt a new framework or credential for hiring outside of a university degree, although there are some successful examples. For example, Multiverse has established corporate apprenticeships as an alternative pathway into work. Many of the 2nd gen companies (e.g., boot camps) also struggled to land on a commercially viable and socially palatable business model. Their primary product wasn’t software; it was white-glove education and training where your CAC is typically high and gross margins are lower than traditional SaaS. You’re also not relying on recurring revenue, which is tough. An added layer of tension was figuring out how to monetize from the candidate side of the ecosystem, which on occasion led to predatory financial products taking advantage of vulnerable candidates (e.g., some forms of ISAs).
Sar: What sticks out to me is the idea that friction is not necessarily harmful. The intentional use of friction is a big theme in how you have designed what you do. I want to get to that later. As far as creating new credentials go, I agree it’s an uphill battle because previous attempts have either failed or paid too little attention to thinking of it as an employer-side problem. New credentials don’t mean anything if the people in charge of reading those signals resist or ignore them. Talk about how you guys are tackling the signaling piece.
Tom: Our perspective at Forage is that employers are uniquely positioned to educate and train candidates, and you can use software to do that at scale. If employers are to trust a credential outside of a university degree, it will most likely be a credential authored by that employer. Rather than reinventing the education system, we’ve focused on building the connective tissue between the world of learning and the world of work by allowing companies to create these last-mile training credentials. And importantly, we found a model where employers foot the bill, given they use our software to combine education with tools to help them build their employer brand and manage their candidate pipeline.
Sar: Give us a quick overview of what you do.
Tom: Forage helps motivated students get great jobs. We do this by working with the world’s top companies (e.g., Goldman Sachs, BCG, Lyft, Lululemon, JPMorgan) to create company-specific online job simulations for candidates to build career skills, confidence, and visibility. The simulations, which typically take between 5 and 10 hours to complete, contain hypothetical life-like projects which mimic what it’s like to work in a specific role at a company. Suppose you enroll in our Electronic Arts job simulation, for example. In that simulation, you learn how to build a game, which includes learning to use languages such as C++ and Python while also learning to minimize cybersecurity risks as you launch that game. Our job simulations do not replicate what you learned at school to give candidates an immersive, realistic, and interactive opportunity to road test different careers. Our partner employers use our software to surface high-intent students while giving them the tools to build meaningful relationships with these candidates long before the recruitment process begins.
Sar: If we look at this market through the lens of getting better at the matching problem, you have taken the evolution of products in your market to its logical conclusion. Why try to get better at matching by being a middleman when both sides can self-select and communicate their preferences? That is like the recruiting version of dating apps! And that is what I thought when I first learned about your Series A in 2020!
Tom: It’s not revolutionary to say the recruitment process is broken, but it is. The solution to fixing this mess isn’t to make incremental improvements to the job matching process - we’ve tried that for the last 25 years and have had little to no progress. Employers need to rethink how they hire, and we believe that starts by recognizing that the recruitment process is currently the wrong way around. Rather than ‘hire then train,’ employers should use a ‘train then hire’ model (especially in the age of software). By doing this, you reduce the information asymmetry which currently exists between employers and candidates, particularly in the early-talent space. On the student side, this flipped recruitment model allows candidates to make more informed and deliberate career decisions by road testing a bunch of careers before committing to one. On the other side, this model gives employers higher fidelity signals to identify candidates who are genuinely engaged and intentional about working at their organization.
Sar: I love framing the “train then hire” model. As we discussed earlier, the second generation of companies also has a version. What’s different is employers drive the model in your case. I want to talk about some principles that are pretty evident in your strategy. The first is the idea of access. Not everyone knows what they want to do or knows people who could show or tell them what a particular job feels like. There’s a lot of education to be done to help people explore. You seem to have optimized for that in two ways. First, you work with employers across sectors and functions. You are forgoing the benefits of specialization around an industry or skill. Second, your courses are project-based and closely proximate to the actual job. Can you expand your thinking on the idea of access?
Tom: When we chatted to students in Forage’s early days (before we landed on the idea of job simulations), we kept hearing this frustration that students couldn’t figure out at a granular level what different careers entailed. The typical refrain was, “If I’m sitting at my desk in this job, what will I be doing?” Students were tired of polished corporate employer branding campaigns; they just wanted a realistic depiction of the role. The idea behind our job simulations is to allow students to road test different careers and get a sense of the type of work they would be doing by actually getting their hands dirty.
It might sound trite, but we want to create a fairer recruitment system by giving all candidates access to this knowledge. We tend to underestimate just how unfair these systems are. Too many employers index ‘elite’ education in the hiring process, even though the data shows that access to an elite education is dictated by wealth. Given the employers we work with hold the keys to economic prosperity, we’ve built a talent attraction product for them with ‘access’ at its core. Any candidate can access our job simulations - there are no barriers such as cost or level of education.
However, some tricky product challenges have emerged regarding access. Internally, we talk about a spectrum of career awareness, with ten being the highest awareness of your desired career path and one being next to no awareness or understanding. We aim to help every candidate get to at least a 7 out of 10 on that career awareness spectrum. But every candidate starts on a different part of the spectrum. If you only have content that helps a candidate at 5 out of 10 get to a 7 out of 10, then you aren’t providing equal access. You’re arguably exacerbating the gap between different cohorts of candidates. The content that works for a candidate at five will differ from that for a candidate at three. We’ve learned that creating truly equitable access is not merely making a bunch of one-size-fits-all content available to everyone. The point is to develop a spectrum of helpful content for everyone, not just a few. So we’ve had to think carefully about how to stitch together job simulation content that allows all candidates to jump on the spectrum at the point that reflects their current capability. That isn’t easy, and we’re still trying to figure out how to do this.
Sar: I love the concept of the awareness spectrum. Aspiring to help people based on their knowledge level makes sense. What are some early attempts to get that right at scale? What is the profile of job seekers?
Tom: When we started building Forage, I thought our simulations would resonate most with college juniors and seniors. I over-indexed my own experience where I treated finding a career as an afterthought. I scrambled to find something I was interested in as I approached graduation (and unsurprisingly, I ended up making a wrong career decision!) Interestingly, and contrary to my experience, our simulations are most popular with first and second-year students, which I didn’t expect. Students are thinking about careers earlier than ever, which is now Forage’s sweet spot. We’ve even found that high school students enroll in our simulations, which has led us to create high school-focused simulations with companies, including SAP.
The challenge with scaling career awareness knowledge (or any education material) is the perceived tradeoff between access and customization. If you’re optimizing for scale and access, that has typically come at the expense of customized content and learning experiences. We try not to see access and customization as mutually exclusive, although reconciling the two can be difficult. We’ve historically prioritized access which has allowed us to pre-skill over 2 million students in a short period. But over the past few months, we’ve started distributing content at different altitudes to cater to a broader audience (going to the career awareness spectrum point raised earlier). As you might imagine, the most important lesson has been looking closely at the data to ensure the product isn’t exacerbating the gap between different cohorts of candidates. We do a lot of work capturing a high-fidelity picture of a candidate pre and post-simulation. That allows us to measure the impact of the simulation on the candidate and where simulation content is and isn’t working. We also look closely at educational and demographic data to ensure the content isn’t leaving anyone behind.
Sar: You mentioned you don’t charge the job seekers. They don’t even have to apply to take a course. Anyone can simply start any course they like. There’s no friction in trying you out. The high drop-off rates in online classes are a big problem for the first and second-generation companies we discussed earlier. It is less of a problem for you because drop-off rates help with quality checks on the employer side. Employers have to only care about the candidates that got the certificates. And, those who complete the course are not doing it to get a bang for their buck! It's all free! So it's a more robust signal of people caring enough about the job. Am I thinking about the filtering correctly here?
Tom: Yes, that’s right. Going back to what I said earlier, the biggest challenge many of our partner employers face is figuring out who is intentionally interested in a career at their company rather than those who are just ‘spraying and praying. In that sense, we haven’t traditionally taken an interventionist approach to job matching (e.g., suggesting candidates apply to certain roles or companies hire specific candidates). Rather, we’re focused on giving candidates the resources and tools to make informed and deliberate career decisions while surfacing the high-intent candidates to employers. What’s remarkable is that 2 in 3 candidates change their minds about where they want to work after completing job simulations on Forage. Isn’t that an incredible indictment of how the current system works? You can spend hundreds of thousands of dollars on a university degree while accessing a crazy amount of online career information. Still, 2 out of 3 candidates change their career decisions after doing a couple of 5-hour simulations on Forage. And we now have had statistically significant volume through our simulations (almost 3 million enrolments), so that statistic has genuine credibility. So often, we hear students come onto Forage saying, ‘All I have ever wanted to be is a banker at Goldman Sachs.’ Then they start doing Goldman’s IB job simulation and realize they would hate doing that job, so they bounce before finishing the simulation. In our view, that’s a great result. The candidate figured out the role wasn’t for them, and Goldmans don’t have to deal with another applicant who is poorly suited to the opportunity.
Sar: In some sense, you are a content business. What’s the scope of design, production, and quality control between you and the companies? You had to go through an evolution in how you approached content after facing some challenges in the old model.
Tom: It might not be trendy to admit this in a SaaS-dominated world, but the content is our product. Forage’s product prepares candidates for the world of work and helps them land jobs that align with their skills and interests. We can only deliver on this if we create meaningful, impactful, and accessible content. Forage is a house of cards if we have bad content.
We’ve had an interesting, at times painful, journey figuring out how to create quality content at scale. Before our Series A, we asked our partner companies to create content themselves. We didn’t want to produce content in-house because it would blow our margins and widen our failure surface area. But getting companies to develop content ended up being even more painful. Firstly, it was a blocker to sales. Companies (rightfully) felt it was ridiculous they would have to create the product they were paying for, which made the sales process challenging. They also didn’t feel they had the bandwidth time to create the content, which became a constant objection in the sales process. You can imagine how hard it was to convince a team of recruiters at an organization like Citi to go and get a bunch of analysts to create the content from scratch! Secondly, employers insisted on owning IP when they created the content. We needed to own the IP. Thirdly, if we did manage to make the sale, the implementation period would drag on for months, given recruiters were herding various people in different business lines to create the content. That created revenue recognition headaches and hurt our LTV, given the subscription start date would start when the simulation went live (not when they signed the contract). In some instances, the first term of these ‘annual contracts’ would last for +18 months, given the long content creation process. Finally, and most importantly, the final product was never great because most companies don’t know how to create content that resonates with students.
While we still work closely with our partner companies developing the content of the simulations, we have now moved to a model where we use independent contractors to do the heavy lifting of creating content using our design methodology. These contractors have previously worked in the specific roles we’re showcasing in the simulations. They know intimately what those roles entail (e.g., if we’re doing an audit simulation with a Big 4 firm, our content contractor must have previously been an auditor at one of those firms). Our content process starts with collecting insights and information from employees in the relevant business lines of our partner companies. Our content team then works with our contractors to design the initial draft of the content before handing it over to the company to review. The partner companies provide input on whether the work accurately reflects what they do, and unsurprisingly, the more effort they put in at this stage, the better the content ends up being. Our team also works closely with our partner companies’ branding teams, given we want the simulation to look and feel distinct to the company (and they can be very particular about their brand!) Once this has all been done, we launch the simulation to our students and university partners.
However, content development doesn’t end at launch. We monitor how the simulation performs relative to our benchmarks, including enrolling, completion rates, and quantitative and qualitative feedback from students. Based on this feedback, we continue to tweak the content - you can never be 100% confident about what does and doesn’t resonate until the content is in students’ hands.
Sar: Can you talk more about how you think about the IP? Does that become a sticking point in closing sales?
Tom: Given content is our product, it’s critically important we own that content. This can be tricky when working with large global enterprise customers who default to overreaching on commercial and legal terms, including relating to IP. But given:
We’ve significantly invested in and built a methodology on what makes excellent, accessible, and impactful simulations; and
We use our contractors to create that content, with our content ops team owning the majority of the content implementation process,
It makes sense for Forage to own the content. It is so important to us that we could probably reduce our average deal close time by 20% if we didn’t insist on owning IP, but the tradeoff is not worth the reduction. Owning the IP also protects against the risk that we spend all this time and effort developing the content only for the customer to post it elsewhere.
Sar: How is the content team structured? Can you share what you think about content production’s quality and scalability?
Tom: Our content team has two dimensions - Content Operations and Content R&D. Content Operations is responsible for content production. This team works with our customers and pool of independent contractors to create the job simulation content posted on Forage. The Content R&D team sits within our product team. Their remit is to understand what good content looks like and build frameworks that allow us to create that content at scale.
We think about content scaling across a brand rather than across an industry. When we develop content with a customer, they are granted an exclusive license to use that content for the duration of the agreement (i.e., we can’t reuse that content with other customers). Our litmus test is that the content should be so bespoke to a brand that you couldn’t reuse it with another brand even if you wanted to. For example, in our JPMorgan software engineering simulation, students contribute to JPM’s unique open-source framework. For example, that content couldn’t be used with Morgan Stanley because they don’t use that framework. That might not sound like a great way to scale content, but we make up for it by scaling that content across a brand’s geographic footprint. Given we work with the world’s largest customers, we can cross-sell into our customers’ offices abroad. For example, we now work with KPMG in the US, Canada, India, UK, and Australia. The content we create for the US team translates relatively well into those international practices, although we do spend time localizing that content. Since KPMG has offices in every corner of the world, our goal would be to sell one data science course to each office, albeit localized for each region. The unit economics of our business becomes increasingly attractive the more you can cross-sell the same content across those different regions.
The other point is that by not duplicating content within an industry, we can expand the surface area of content and skills students can develop. For example, we work with several financial institutions (e.g., HSBC, Goldman Sachs, Citi, HSBC, and Bank of America). We worked with these customers and industry experts to map out the 253 skills (both technical and soft) required to be an investment banker. We then divide those skills into bundles, which we sell to customers. This means we’re not replicating content for every new investment banking simulation we bring on board but rather allowing students to go deep and develop their skills.
Sar: Geography as a vector for scalability across employers is very interesting. Honestly, I didn’t even consider that when I was thinking through your business. How you approach content necessitates having a particular employer persona. You told me that you have a growing presence in India, and many of the customers are Indian subsidiaries of American companies. Walk us through the localization dynamics as you enter new markets.
Tom: From day one, we decided to go global. Besides Pasha (my co-founder), everyone said going global from the start was a bad idea. The advice was to focus on the Australian market and do that well. But given we had limited capital, the Australian market was too small to grow revenue quickly and meaningfully. We decided to target the offices of prominent global brands in Australia (e.g., KPMG, Citi, Accenture) and then cross-sell into their international offices. The international sale was more straightforward because their Australian peers (our customers) could vouch for their experience with our product. These were those against-the-grain decisions early on, which were crucial to our survival and growth.
Today, we primarily focus on selling to global brands in the US and then look to cross-pollinate across different business units and geographies quickly. We now work with North America, Asia, Africa, Europe, and Australasia companies. We can do this because the problem we’re solving transcends culture, geography, and language. Universally, there is a significant gap between the world of learning and work, which is why we have customers worldwide. But, that gap does look different in each of those markets. The three vectors to localization for our simulations are:
Audience: The audience for our simulations can change depending on the geography, which can impact how the content is designed. For example, consider students interested in a career in law. In the US, you need to do a JD to become an Attorney, meaning you’ve already completed undergrad and most likely have had career experience. That’s very different from the UK, where you don’t need to study law at university to become a solicitor. The content you create for a cohort of primarily +24 year-olds with a degree and work experience is very different from the content you would create for 18-year-olds.
Distribution: Universities are one of our largest distribution engines, and this engine differs significantly between different markets. Ideally, we get universities to integrate our simulations into their curriculums so that professors assign the content as coursework. However, this is far easier in some regions compared to others. In the UK, for example, curriculums are heavily influenced at a faculty level. If you can build a partnership with a faculty, all students in that faculty will use our content in their coursework. In the US, however, curriculums are more heavily influenced at a professor level. So you could expend the same energy trying to get a professor to integrate in the US, which would generate 40 odd enrollments, as with a faculty in the UK, which would generate 1000s.
Content: Finally, we must ensure that the content is relevant to the specific geography. We have had instances where the content has missed the mark in different markets. For example, we had a simulation in India with a professional services company which we had to re-do because the client subject in the simulation was a gaming company. We adapted this simulation from a simulation in another country. The customer was frustrated, given it was very rare for an Indian consulting company to work with a gaming company.
Sar: Talk more about the persona of the companies you work with. The segment you are going after has implications for your top-of-the-funnel growth amongst students and the company brand.
Tom: Our ideal customer profile is large global brands who do high-volume entry-level recruiting (i.e.,> 500 annual entry-level hires) across multiple business divisions and predominantly hire out of 4-year college degrees. Initially, we target the US offices of these brands before cross-selling into other regions. To give you a sense, our existing customers include Walmart, Accenture, KPMG, PwC, SAP, Dell, and others. There are a couple of reasons why this is our ideal company persona. Firstly, our early insight was to target large recognizable brands. We knew that we couldn't kickstart the candidate side of the marketplace with the Forage brand, given no one knew who we were. But we could piggyback off our customers, who had more brand equity than we could ever achieve. Once we could get these brands on, it became insanely easier to attract students. Secondly, these large brands more sharply experience the problem we’re solving. They are sending hundreds of campus recruiters to schools trying to hire thousands of candidates. They spend millions of dollars on the process only to experience 70% attrition the first three years of that hiring class starting. I once asked a head of early talent how he thinks about solving the challenge of 70% attrition. He responded that their solution was to hire 70% more candidates. That’s how hopeless the situation has become. Finally, these companies have the budget to pay for higher-ticket HR software.
Sar: Can you share your thinking on how GTM works for you? Do you work with different teams during the sales and post-sales process? What is the onboarding cycle like? Most of your competitors sell your customers on either lead gen or applicant tracking systems. Do you sell both? What are some challenges and benefits of having to sell an integrated service?
Tom: We work with three distinct groups of people during and after the sales process. The first group is the executive sponsors. These tend to be C-suite executives or senior leaders in different business divisions. Having an executive sponsor to get a deal done is not a precondition, but it makes it much easier. The second group is our economic buyers. A usual role in this category is VP of Talent Acquisition or Director of Recruiting. They own the budget we sell into and will be a crucial part of any sales process. While this group doesn't use our product, they are the group that holds us accountable for the value we say it delivers. The final group is the users of our product. They are the teams of recruiters sourcing candidates who tend to be earlier in their careers. Building deep relationships and showing value to each distinct group is crucial to the success of our sales and expansion engine. However, in the sales process, we primarily work with that second bucket - the economic buyer. It’s not until post-sale that we work closely with the third group, but they become critical regarding renewals and expansion.
Sar: That sounds like quite a challenging process involving so many parties.
Tom: It’s not easy selling what we’re selling. We’re building a category-defining product. I don’t say that because it’s glamorous, because the reality is it isn’t. I deeply admire other companies that have successfully built category-defining products in our space, including Guild, Multiverse, Handshake, and Degreed. But the challenge with selling a category-defining product is it is rarely a sub-in / sub-out product, unlike an ATS or CRM tool. Prospects will always default to trying to find an existing analogous product. But our job is to educate a customer on why and how the future of recruitment will look different and why our brand new tool will help them succeed in that new world. For us, that future is characterized by the ‘train then hire’ model and our ambition is that Forage is at the forefront of that new world by enabling companies to meaningfully pre-skill and then hire great talent at scale.
Sar: I want to now talk about the student experience. Let’s say I'm a job seeker who happened to know about your website. I get there, explore a few courses, take a course, and get my credential for completing the course. All of that happens within your product experience. What happens next? You don’t take on the task of assessment, right? What parts of the recruitment process happen on Forage from the candidate’s standpoint? When does the relationship go off of your product?
Tom: Currently, we’re only focused on the career discovery space (and don’t act as an assessment tool). Our website is similar to a Coursera or Netflix experience. You jump onto Forage, browse through our catalog of simulations, and then pick and choose the simulations you want to complete. Once you complete a simulation, you can use that experience on your CV, in interviews, cover letters, etc. We also direct candidates to our partner companies’ recruitment campaigns (and software) after they complete simulations, whether to their career pages or ATSs. Historically, Forage has played a non-interventionist approach to helping students secure employment (beyond creating meaningful simulations). But that’s starting to change. We are now experimenting with wrap-around services to support students, including testing cohort-based classes focused on preparing students for the job application process. And even though we aren’t an assessment tool, we are increasingly surfacing high-intent candidates in our back-end software to our partner companies.
Sar: Companies manage first-party/third-party job listings and various application tracking systems when we think about inbound and outbound recruiting efforts. Where does Forage fit amongst the wide range of vendors and applications they might use? Do you think of your employer-facing system as an integrated lead gen and management system, or are there plans to help consolidate many applications your customers might be using over time? Do you integrate with other systems?
Tom: Recruitment is a similar function and muscle to sales. Despite this, recruitment teams tend to do poorly at taking a data-driven approach to nurturing their top-of-funnel candidate pool. Companies invest in employer branding but are (bluntly!) flying blind to understand the top of the funnel intimately. There are a few reasons which sit behind recruiter inertia on this. Firstly, employers haven’t typically seen the value of nurturing a talent pipeline before the recruitment process. There are a handful of candidate CRM providers, including Gem and Beamery. But that leads us to the second problem - there haven’t been strong incentives for candidates to enter these top-of-funnel ecosystems. If you aren’t actively looking for a role, chances are you’re not likely to go out of your way to opt-in to a company’s CRM ecosystem. Given this, most of our customers invest heavily in building a software stack that starts at the point of a student applying for a role (i.e., Applicant Tracking Systems and beyond).
But things are changing as the competition for talent continues to heat up. When students reach their junior year, the race to win that talent is over. In that context, employers realize they need to engage, understand and nurture early talent earlier than ever to win. Our product lets employers do this; critically, we’ve built a product that incentivizes students to participate in this ecosystem even though they are far from applying for a role. Our back-end software is similar to other candidate CRM tools, but the key difference is the volume of candidates in that tool is significantly larger. The only way to get students into this ecosystem is to deliver genuine value on the front end. The companies that win talent in the future are the ones who give more than they take in the recruitment process, and job simulations allow them to do that.
Most HR tech companies will tell you that hiring attribution (i.e., tracking hiring outcomes of their candidates) is the bane of their existence! Even the giants in our space have struggled to solve hiring attribution, most notably Indeed. But the best way to track hiring attribution is through CRM and ATS integrations. ATS integrations allow our customers to track how many candidates are sourced from our simulations easily. This data is also critical for us to track the effectiveness of our product and whether the simulations lead to great employment outcomes. Unfortunately, our customers all use different ATSs, and only some ATSs are open to integrations. Many legacy ATSs are deeply resistant to integrations, whereas some of the newer players (e.g., Greenhouse) actively encourage integrations (which we love to see). Without these integrations, we have other imprecise solutions, including tools in our back-end software that enable employers to cross-reference hires and self-reported data from students and companies. Given that we work with large enterprises with close relationships with those organizations, it’s not as tricky as HR vendors dealing with thousands of SMEs.
Sar: Can you walk us through your business model and how it ties into your growth strategy?
Tom: Forage charges employers a fixed annual subscription fee to post job simulations on Forage. We also charge an implementation fee which covers the cost of creating the content and managing the implementation period. As part of the subscription, our customers access HR-enterprise software (discussed above), enabling recruiting teams to manage their candidate pipeline and identify high-intent candidates. The subscription fee increases when the company decides to post more simulations, whether that’s across other business divisions or geographies. In this sense, we use a ‘land and expand’ model - we try to land somewhere small, to begin with and then rapidly expand that account once we’re through procurement and vendor onboarding (typically the most painful part of the sales process). In our experience, it can spook a prospect if you try to sell too many simulations during the initial sales process, which is why we land small initially and then expand once we have proven our value with a low-risk trial.
We haven’t experimented and innovated enough on pricing. It’s an area most startups gloss over, and we certainly fall into that bucket. I’ve recently reflected that internal factors drive our pricing structure on what makes a good business (e.g., how do we price for strong gross margins, CAC payback period, etc.) rather than the actual value we deliver to a customer. My inkling is we leave money on the table if you look at the value our product provides to a customer. If you were to divide the subscription fee our customers pay by the number of hires we generate for them, that cost paid per hire number is very low compared to industry standards, indicating there is room to grow.
Sar: Ah, you are saying you are underpricing your offering but are still doing well on economics because the pricing was designed keeping internal factors around economics in mind? Have you had other ideas for upselling the geographies or business units of the customers you already serve? I’m also wondering who you consider “customers”? You referred to employers as partners earlier.
Tom: Our ‘customers’ are the companies that pay for our software (I was perhaps sloppy with my language earlier!). If you look at the ‘cost per hire’ customers are paying for Forage hires, they are getting a great deal! But yes, we’ve managed to keep unit economics attractive despite leaving value on the table.
In terms of upselling, we’re focused on breaking into enterprise with our job simulation product, to begin with. The enterprise sales motion is hard - you have to get an initial break-in, cultivate an internal champion, get a buyer on board, get the timing right, and hope there’s enough budget. But the part not often talked about is getting through legal and procurement - this can be the biggest timesuck in the process. However, a difficult sales process reward is that enterprises can be very sticky once you are onboarded. Given the pain the customer goes through to onboard a vendor, their initial preference for buying new software is from the companies they already work with. You see so many enterprise customers with incredible net revenue retention rates.
At some point, we will diversify our product offerings and make the most of these enterprise partnerships. It’s a matter of when not if. We’ve already experimented with selling different products, including ATS software, assessment solutions, and software that enables companies to run their internships remotely. While we’ve not put the required effort to get these products truly enterprise-ready, we’ve proven we can sell these adjacent products to our customers. To avoid the ‘distraction’ trap, we’re focused on breaking into enterprises with job simulations for the time being.
Sar: Given where you sit in the value chain, you are positioned to take on workforce upskilling experience for your customers. Do you guys have any plans on that front? What do the next two years look like for you?
Tom: It’s easy to lose focus in our space - so many different directions we can (and do) get pulled. There have been instances where we lost that focus with distractions outside the pre-skilling space. But our mission is to level the recruitment playing field, so we want to remain focused on the pre-hire space. Right now, we focus on the early-talent space and will continue to do so for the next two years. That’s because there is sharp information asymmetry between candidates and employers. They do a poor job of understanding one another. But the ‘train then hire’ model makes sense wherever information asymmetry exists between a candidate and a role. Early talent isn’t the only place this works. Our customers ask for simulations designed for military personnel transitioning into civilian life, parents returning to the workforce, and staff looking for new careers within an existing employer. Our ambition is that customers will eventually deploy simulations across different business divisions and regions for candidates at different points of their career journey.
Sar: Can you share one such distraction you might have had in the past?
Tom: That is a hard question because I’ve thought about whether these distractions were necessary to our journey. Perhaps we will always get distracted at some point, and you don’t learn the lesson until we make a mistake. Regardless, the obvious distraction is we fell into the trap that many early-stage HR companies fall into - building an Application Tracking System. If there is one thing I’ve learned, the world doesn’t need another ATS. The ATS provider of one of our early customers experienced a data leak which created bad publicity. In light of that, one of our early customers wanted to change ATS providers and asked whether Forage would become their ATS provider. At this stage, we were sub $1m in ARR, and the opportunity seemed too good to pass up, given the ticket size of ATSs compared to our product. In just a few weeks, my co-founder and our only other engineer built an ATS that could process a recruitment campaign of over 3000 applications. It was a big win at the time, but the ongoing maintenance and problems it created far outweighed the revenue we generated from it. ATSs is complicated software to build and a highly saturated space. In saying this, I still believe it’s hard to build a +$100m ARR business in the HR space without tackling a few different products. However, you need to make a judgment call on when it makes sense to pull the trigger on building your next suite of products after your core product.
Sar: Let’s switch gears a bit. You started the company and raised a pre-seed round in Australia. You later went through YC and moved to the States. Your cofounder is still based back home. You believe that Australian founders are more comfortable in this new capital environment than we have had for the past five years. Can you say more about that?
Tom: The Australian start-up ecosystem was nascent when we founded the company in 2017. Atlassian was the only true homegrown success (although Canva and a few others were on their way). Even today, the availability of early-stage capital in Australia pales compared to the US. Forage (like most companies) found it challenging to raise pre-seed money in that environment and struggled to even set up conversations with local funds. We raised just US$30k in pre-seed capital from an Australian accelerator, which we lived off for the company’s first year (we’ve now raised $38m for context). That first year bred a sense of extreme capital consciousness - every cent was scrutinized, and we had no choice but to figure out how to use it to make meaningful revenue from our customers. Given that starting point, I’ve always felt more comfortable in an environment where you have to figure out how to go further with less. Having raised our Series A and B during the craziness of the last two years, I counterintuitively found that environment more challenging where the expectation was to deploy capital aggressively to pursue top-line growth. I know many international founders who share similar feelings. In saying that, I certainly have a sense of apprehension about the next few years but strongly believe that companies with great fundamentals are born during fundraising winters, given there is no other way to survive. We intend to be one of those companies.
Sar: What do you want Forage’s legacy to be, and what are you grateful for these days?
Tom: We often hear about the challenges of being a founder, including anxiety, loneliness, and financial pressure (particularly in the early stages). But my greatest career privilege is the opportunity to solve a problem you care deeply about with wickedly smart and passionate colleagues. A friend recently asked me what I intend to do after Forage. The truth is, I have never thought about it. I’m not sure many people can genuinely say they are so satisfied in their role that they wouldn’t be tempted to swap it for another role. But I am one of those people, and I try my best not to take that for granted. Outside of work, I am also incredibly grateful to be a father. I founded Forage before becoming a dad, and it’s fair to say that an unhealthy amount of my identity was tied up in the company. Every second of every day was consumed thinking about the company, and my happiness was inextricably tied to the company’s success. Of course, Forage still does consume me, but becoming a dad has given me an identity I love outside of being a founder. There is no better way to disconnect than taking my son Rory to the park.
When it comes to Forage’s legacy, I challenge our team to internalize the ‘but for” test. ‘But for’ Forage, would this first-gen student from this overlooked university have landed this fantastic role with JPMorgan? I don’t want a correlation between candidates who do Forage’s simulations and candidates who achieve great employment outcomes; I want Forage to be the cause. But ultimately, we want to build a product that allows companies to build better workforces. For too long, HR vendors and recruiting teams have only truly focused on hiring efficacy. That’s not the point of recruitment. The point is to attract great people who love their jobs and make meaningful contributions over a sustained period. For Forage to reach its potential, that causal connection needs to extend to Forage, ultimately creating a better workforce.
Sar: What are you bearish and bullish on in the long term?
In the long term, I’m bullish that we’ll move to a system with a tapestry of different pathways into entry-level professional employment. Those pathways will include corporate apprenticeships, boot camps, and employer-created credentials. Right now, corporate America has an unhealthy addiction to hiring people with a 4-year college degree. It’s not effective, and the data shows it. My caveat is we won’t get to this new system because employers recognize how ineffective sourcing from the 4-year college degree is, but rather, they will have no other choice. We already see this with in-demand jobs like software engineering, data science, and cybersecurity. Employers, including AWS and IBM, work more closely with community colleges and other institutions outside four-year colleges.
The inverse of this bullish perspective is a bearishness on the future of the university system (at least as we currently know it). Don’t get me wrong; universities are a critical plank in society. But consumers overwhelmingly see the value in the university product as the employment outcomes that flow from it. And too many universities aren’t currently providing that value, so institutions close down at a rate not seen before, and overall enrollments decline. This won’t ever impact the top institutions, but in 20 to 30 years, I suspect the percentage of people obtaining a 4-year college degree will dramatically decline.
Sar: What do you think we are not paying enough attention to? It does not have to do anything with what we talked about
Tom: The fallacy of the American Dream. I love this country, and part of that love is based on being able to call out BS when you see it. I’ve observed a certain dogged attachment to American ideals even where the data tells a different story. One of those ideals is the American Dream - the idea that anyone with the will and drive can live out the “rags to riches” story. We all hear extraordinary anecdotes of the American Dream, but the data shows these stories are the exception, not the norm. The greatest single determinant of whether an American ends up in a high-paying job is the zip code they are born into. It’s quite literally the opposite of a meritocracy. And for some reason, otherwise intelligent people get riled up about this fact. Instead of running towards the fire on how they might solve it, they spend a ton of energy defending a flawed system. The uncomfortable truth is that even if the American Dream initially had credibility (which I’m not sure it did), the data shows upward mobility is regressing. It’s now harder to progress up the economic ladder than 80 years ago (based on data from Opportunity Insights). In 1940, 90% of people earned more than their parents. Today that number is closer to 50%. Isn’t that cooked?! Shouldn’t we look at doing something about this?
Previous chat :
Thanks for reading SarHaribhakti's Newsletter! Subscribe for free to receive new posts and support my work.