In 2018, while I was a PM at Indeed, I came up with an idea to change the way jobseekers found and landed work. Armed with some initial user testing, I applied to the inaugural class of Indeed's internal Incubator program. I was accepted.
Four years later, alongside a small, incredibly talented and empowered team — and with the might of Indeed's sales organization behind us — we had grown that idea into a full-fledged product. By the end of 2022, it had an annualized revenue run rate of more than $300 million.
This is the story of how that came to be.
It started with a Help Wanted sign
The story begins a year earlier, in 2017. The first team I joined at Indeed was a two-person team that had just graduated from Indeed University. The program gave recent college grads an opportunity to explore the market and build whatever they wanted. It was a bit like an internal Y Combinator, and a precursor to what would become Indeed Incubator.
The team had developed a product called JobSpotter. The pitch was simple: aggregate jobs in the physical world, particularly jobs that were traditionally offline — think Help Wanted signs in restaurants, bars and retail stores. This fit neatly into Indeed's founding strategy: aggregate all of the world's jobs into a single search engine. It also brought us into contact with employers who hadn't established a relationship with Indeed and represented a largely unmonetized part of the market.
JobSpotter's approach was novel. We built an app and gamified the process of capturing Help Wanted signs around the country. Take a picture of the sign and the storefront, then submit. Behind the scenes, we captured the GPS location, removed recognizable faces, checked the photos for quality and authenticity, and created a digital job listing on Indeed.
Each submission earned a reward based on rarity, image quality and the type of establishment. We placed more value on mom-and-pop businesses than corporate locations, and paid users in Amazon gift cards. Leaderboards added another incentive. Over the course of a year, users captured more than a million jobs, and Indeed paid out thousands of dollars in gift cards.
By and large, it was a big success. But there was an unintended consequence: Indeed was sometimes too good at generating awareness. We were overwhelming these employers with applicants.
I reached out to business owners to understand what was happening and how their recruiting processes were working. A couple of themes kept coming up. Applicant volume could be a challenge, but the bigger problems were whether candidates were qualified and whether they arrived when someone could actually interview them.
A manager might only be available on Thursdays. A candidate who stopped by on Monday could leave an application, but nothing would happen until Thursday.
More applications weren't necessarily the answer. The opportunity was to get qualified people in front of an available hiring manager.
Breaking down the problem
I thought both challenges could be solved — or at least I had a hypothesis that they could.
Quality. The jobs we were targeting — retail, hospitality, restaurants and similar roles — required a mix of things. Employers assessed communication and other interpersonal skills alongside job requirements such as work eligibility, the ability to lift 30 pounds or the ability to stand for an eight-hour shift.
My hypothesis was that a few screening questions could cover some of those requirements, while an interview would cover what a form couldn't. For these employers, attending an interview in person was also part of the screening process: could the candidate get to the work location?
Timing. Why submit an application on Tuesday if the manager wasn't going to look at it until Thursday? What could we offer jobseekers that would make it worthwhile to align their visit with the manager's availability?
My hypothesis was a simple calendar-based scheduling tool, built around the business's schedule and one question: "When do you want to do interviews?"
The more nuanced follow-up was: "How many people are you looking to hire?"
That sounds simple until you work backward. How many interviews would it take to make those hires? How many RSVPs would it take to get that many people through the door?
This led us to what we called show rate: the number of jobseekers who attended, divided by the number who RSVP'd. Predicting and improving that rate would become one of the core challenges of the product. More on that later.
Proving it out
Okay, so I had a hypothesis. Now I needed to test it.
I built a simple RSVP form in Typeform with a few screening questions. Then I convinced a few colleagues to create a job feed for those forms, much like the feeds Indeed indexed from other sites. That made the events searchable through the core Indeed interface.
I went back to two employers I'd spoken with: a small, single-location bakery in Denver and a McDonald's in Memphis. Both were hiring, and both had motivated managers willing to interview candidates — and make offers on the spot — at a specific time on a specific day.
I configured the forms with each business's screening questions, collected names, emails and phone numbers, and let it fly.
Would anyone see them? Would anyone RSVP? Would anyone show up? Would anyone get hired?
With Indeed's traffic, it didn't take long to answer the first two questions. Jobseekers started signing up. Maybe this thing was working.
But an RSVP wasn't an interview. I still needed to reassure people that these were real opportunities and remind them to attend.
At first, that was completely manual. I emailed every jobseeker who RSVP'd, thanked them, explained the process and reminded them what to bring. The day before the event, and again a few hours before it started, I sent emails and text messages with the time, location and a good-luck message.
I kept my fingers crossed.
The next day, I called the managers. Both businesses loved it. They had made the hires they needed in a single day. The feedback was simple: it couldn't have been easier. The McDonald's was part of a larger franchise group, and they wanted to use it across their network of 30 locations.
Meanwhile, emails started arriving from jobseekers thanking me. Some couldn't believe they'd been hired so quickly. Even people who didn't get hired appreciated knowing where they stood.
That last part stuck with me. One of the most frustrating things about looking for work was applying to hundreds of jobs and hearing nothing back. The better Indeed got at attracting jobseekers and making applications easy, the worse the problem became. Internally, we came to call it "the jobseeker void."
I repeated the process a few more times with similar results. Employers liked it. Jobseekers were satisfied. I was onto something.
Now I needed to figure out how to scale it.
Getting into Incubator
At the same time, Indeed was forming a new program called Incubator. Its goal was to build new products within the company that could create new lines of business and revenue.
The invitation went out to everyone: pitch your idea to the senior leadership team. If selected, you'd get a small team — product and engineering, with shared UX and sales support — and funding to pursue it. It worked like an internal venture fund, with funding gates tied to goals. Keep meeting your goals, and you'd get more funding.
The timing was perfect. I took my research and early test results and developed a pitch: Indeed Hiring Events.
It was accepted!
Lesson learned: An idea is cheap. Validation with actual users is what matters. Test as cheaply and quickly as possible, in a way that resembles the experience you want to build. A survey asking whether someone would use your product can be useful, but it isn't the same as watching them use it. Take the feedback, adjust and test again.
Day 0
Hiring Events was funded, our team was established, and we had a kernel of positive user feedback.
Step one was to get off Typeform. We built a basic RSVP experience that looked and felt like Indeed, then automated the confirmation emails and SMS reminders so we didn't have to send them manually for every signup. We built it on a basic content management system already available to Incubator teams.
Then we went back to our pilot customers, ran more events and kicked off a scrappy sales campaign to find new prospects. Every customer taught us something. We started to understand what show rates to expect for different jobs and locations.
Turning events into a business
Our business model came from Indeed's core business: job advertising.
We worked backward from the employer's hiring goal. How many attendees would they need? How many RSVPs would produce that turnout? How many clicks would it take to generate those RSVPs?
We grouped campaigns around typical hiring needs and the traffic we could reasonably drive. Packages started at $1,500 and $2,500 and went up to around $6,500 per event.
The campaign window wasn't fixed. We configured it around the number of RSVPs an employer needed and the supply of jobseekers in the region.
We took that budget and advertised on Indeed on the customer's behalf, driving traffic to an RSVP page instead of a traditional job page. The job type, location and employer were all factored into the pricing, and we could use the sophistication of Indeed's Sponsored Jobs technology rather than build an advertising system from scratch.
The employer wasn't just buying an RSVP form. They were buying a campaign built around a hiring goal.
As the months went on, we added features that helped us earn more business or addressed something we'd learned from an event. Our small team forced us to focus on what mattered. Recurring check-ins with Incubator leadership and the senior leadership team kept that focus sharp.
Lesson learned: A small team with funding tied to results creates a particular kind of focus. We knew we wouldn't lose our jobs if the idea failed, but we could lose the team and the opportunity to keep building together. Everyone who signed up for Incubator wanted to win. We were committed to the product and to each other.
Scaling the product
Moving beyond organic traffic meant tapping into Sponsored Jobs. There was a lot of sophistication available, but it wasn't a plug-and-play solution.
At least in 2018, Sponsored Jobs was designed to distribute a job's budget in a pattern tuned for traditional online applications: a short burst at the beginning, followed by a slower burn over roughly 30 days. Employers got some applicants quickly, then a steadier flow over the rest of the campaign.
Hiring Events needed something different.
If your event was on Saturday, advertising it 30 days in advance wasn't particularly useful. Jobseekers might forget about it or find another job. Show rates were significantly higher among people who RSVP'd closer to the event date.
But that didn't mean we could wait until the last minute to advertise. The right campaign length depended on how many RSVPs the employer needed and the supply of jobseekers in the region. Too short, and we wouldn't get enough RSVPs. Too long, and we'd spend money bringing in people who were unlikely to show up. Beyond 14 days, campaigns showed diminishing returns because those early RSVPs had very low show rates.
The algorithm had to balance the two: enough time to generate the RSVPs we needed, with more of the advertising concentrated near the event, when an RSVP was more likely to turn into an attendee. We weren't just trying to maximize RSVPs. We were trying to get the turnout the employer actually needed.
That balance became part of one of the most complex and important pieces of the product: getting the right number of qualified candidates to the right place at the right time.
The complexity grew with the events. Helping a bakery hire one person on a Thursday was one thing. Helping a large employer bring in hundreds or thousands of candidates over a weekend was another. As we moved from small businesses into enterprise accounts, we had to watch events closely and forecast whether RSVPs would translate into the attendance and hires customers expected.
And, of course, there were upper limits too.
At one event for a big-box furniture retailer, we drove so many candidates that the retailer had to call the cops. More turnout was not automatically a better outcome. We had another problem to solve: RSVP caps.
Helping customers understand job requirements, screening criteria, event timing, campaign length and the available candidate pool became part of our special sauce. It also gave our sales team the expertise to be consultative rather than simply take an order.
We could help customers shape an event around what was likely to work. That allowed some of the world's largest retailers to recruit at a speed and scale we hadn't previously supported this way on Indeed. Our customers ranged from small businesses to some of the largest employers in the world.
Lesson learned: You won't know every complexity at the beginning. Launch, listen to customers, learn from mistakes and keep improving. For us, learning quickly mattered more than trying to build for every possible scenario up front.
Sales as our third customer
Working with the sales team was just as important as building the product.
That became especially clear as we moved beyond Incubator's dedicated sales support and into Indeed's broader sales organization. We were an upstart competing for attention with established products like Sponsored Jobs: proven, familiar and easier to sell.
We didn't think it had to be either/or. Hiring Events and Sponsored Jobs could work alongside each other, giving reps another way to help their customers and grow revenue.
So we treated Sales as our third customer, alongside jobseekers and employers.
Their needs were different. It wasn't enough for the product to work. It had to be easy to explain, easy to implement and priced appropriately. Reps needed to feel confident recommending it.
In the early days, the biggest concern was setting a customer up for success. The worst outcome was convincing someone to try a new product, only to have nobody show up. We couldn't eliminate that risk, but we could help reps understand it.
We built a sales-facing pricing and show-rate estimator. A rep could enter the job and location, then work through a conversation with the customer: "How many hires do you want to make?" "How many interviews does it typically take to make a hire?"
From there, we could estimate the RSVPs and budget required, and sometimes recommend multiple events. Sometimes the answer was that Hiring Events wasn't a good fit. We considered that a good outcome — much better than taking someone's money and failing to deliver value.
Beyond the tools, a lot of it came down to being present. We gave reps the same hands-on support we gave our other customers, from starting campaigns to responding to questions and requests.
We developed training materials, ran training sessions and hosted Q&As where reps could suggest features and improvements. We encouraged successful sales leaders to share their approaches with peers. We met regularly with sales leadership and acted on their feedback.
We also made sure Hiring Events was understood as an adjacent product, not a replacement for Sponsored Jobs. That mattered: reps still had large quotas tied to the core product. We needed to help them succeed, not become a distraction.
Over time, we became a trusted partner.
Ultimately, it worked. Our ad hoc approach became the playbook for launching future products through Indeed's sales organization. Hiring Events also became the highest-revenue-generating product per employee in the company's history at that time.
When we passed $100 million in annualized revenue run rate, our team was just 10 people across engineering, product, sales support and design. That count didn't include the broader Indeed sales organization that helped us get there.
Lesson learned: Don't treat Sales as a generic channel. Reps are pitching your product every day and hearing feedback at a volume you may never match through your own outreach. Learn from them. And remember: when products compete for attention, working well isn't enough. They also need to be easy to sell.
The first virtual experiment
Sometime in 2019, we started experimenting with Virtual Hiring Events.
The idea seemed obvious. Why did jobseekers need to travel to a location to interview? Could we bring the same experience online? Remote jobs were a small but growing category, and plenty of office jobs already used phone or video calls for a first interview.
We built a prototype that let employers turn on a virtual option and add a video conferencing link, such as Zoom or Teams.
The response was lukewarm. We got a few bites, but nothing really stuck.
Some of the feedback went back to why in-person events worked for our existing customers: getting to the interview was part of the screening process. But there were product problems too.
We couldn't measure show rates directly, so we had to ask employers and jobseekers afterward. The patterns also didn't match what we'd learned from in-person events.
Candidates often arrived at the beginning of an event rather than at their RSVP time. We couldn't manage their entry, tell them how long the queue was or notify them when an interviewer was ready. Unless an employer used a more sophisticated video platform with multiple rooms and a lobby, they couldn't run separate interviews at the same time.
A video link wasn't enough to recreate the event experience.
We also ran into applications from candidates in other countries and high volumes of unqualified candidates that filled the RSVP slots prematurely. We were facing a familiar Indeed problem: lots of applications, including people who didn't meet the requirements or didn't respond when the employer followed up.
There was probably room to keep improving it. But the complexity was high, and the return on our small team's time wasn't there. We shelved it.
Maybe the world wasn't ready.
Lesson learned: Every expansion should face the same scrutiny as the original product. Don't get drunk on success and assume you can keep expanding, especially when there's still work to do on what's already working. Shelving an idea can be the disciplined choice.
March 2020: The world changes
Then COVID hit.
Offices closed. Retail closed. Hiring plans went on hold. At first, everything seemed to freeze. Indeed's revenue was in free fall, and the company didn't have an obvious response.
As businesses started adjusting, hiring needs returned — but in-person contact was still restricted. Essential workers were in demand across healthcare, emergency services, food retail and services, transportation, childcare and social services.
These were categories where Hiring Events had worked well. The catch was that employers couldn't run interviews the way they had before.
But we had something on the shelf: Virtual Hiring Events.
The need had changed dramatically. Now we had to fix the parts of the product that hadn't worked the first time.
We built a browser-based video conferencing experience using Twilio's SDKs, paired with a virtual waiting room, queuing, messaging and alerts. Employers could run separate interviews in parallel, while candidates waited in a managed queue and received updates. We could also measure attendance directly, giving us the visibility we needed to model show rates and advertising budgets more accurately.
This made larger virtual events possible. It wasn't just a matter of turning the old prototype back on. We had to build the missing pieces that made it usable at scale.
During those early months of COVID, Hiring Events became the only Indeed product growing in revenue. The company leaned into it, supporting major campaigns to recruit essential workers, including a nationwide event in the UK to hire NHS workers.
At first, all of our events were virtual. As COVID restrictions eased, I'd estimate that around 45% of our events remained virtual, with the rest held in person. The move into virtual events accelerated our expansion into more countries and new job categories.
Lesson learned: Timing matters, but it isn't the whole explanation. Sometimes a market isn't ready. Sometimes the product is missing something essential. Looking back, our virtual experiment may have had both problems. The lesson was to revisit the opportunity when circumstances changed, without assuming the old product was suddenly good enough.
When Hiring Events became a platform
Virtual Hiring Events raised the profile of the product and the team within Indeed. The CEO at the time called it one of the company's most important products, and set an ambitious company-wide goal: 10 million interviews in a year, up from our baseline of around 700,000.
At first, this was exciting. Our vision had the spotlight. We could finally scale it in the ways we'd hoped.
The challenge was that nobody had a clear path to that goal. It represented a huge increase in scale, with a lot we still didn't understand.
Leaders from across the company rushed in, each trying to adapt the product in ways they knew how. We developed features that attempted to broaden the definition of an interview to include a one-minute phone call. From my perspective, we were starting to optimize for the goal rather than the customer problem that had made the product work.
The name changed from the simple, descriptive Hiring Events to the broader Indeed Hiring Platform. We changed the pricing and added a platform fee. Customers pushed back, telling us we were charging more without delivering more value.
We lost focus and tried to do too much. Hiring Events was compelling, but it wasn't designed to solve every hiring problem or work for every type of job. We spent more time on media engagements than meeting customers and refining the product.
And I got caught up in it too.
I had a seat at the table with senior leaders that I hadn't had before. I now led a huge team with multiple layers of management. Of course I wanted the product I'd helped build to get more attention. I also enjoyed my elevated position in the company.
I did push back on strategy decisions, particularly the platform fee, which I believed was the wrong way to raise prices. But I also lost sight of some of the principles we'd used to grow the product in the first place.
The message became muddled. Indeed Hiring Platform started to mean everything Indeed did, rather than a clearly defined product. Our growth rate plateaued, and the novelty of Virtual Hiring Events faded among the other products the company was promoting.
That doesn't erase what we accomplished during that period. We helped millions of jobseekers get jobs, and the product reached a scale and level of recognition it hadn't had before.
Eventually, the dust settled. The product returned to the Hiring Events name and entered a more mature phase, with a steadier level of revenue alongside Sponsored Jobs and Indeed's other core products.
Lesson learned: Ambitious goals can force you to think differently and find improvements you wouldn't otherwise pursue. But aggressive goals need clear definitions and governance. Otherwise, hitting the number can become more important than moving the original vision forward.
An unrealized vision
I'm extremely proud of what our small, talented team achieved, alongside a world-class sales organization. I learned an enormous amount during those years. I got to operate like a founder, with real agency and autonomy and a clear focus on creating value for jobseekers and employers.
But we didn't fully realize the original vision while I was working on it.
We imagined a job search engine that wasn't just a list of static jobs. It would show actual interview times and locations that qualified candidates could book themselves. More sophisticated screening and dynamic scheduling could replace the traditional application process with something more direct: answer a few questions, understand whether you meet the requirements and schedule an interview.
For the jobs suited to that model, the jobseeker void could have a solution. Initial screening could happen before the interview. Time to hire could shrink from months to potentially hours. Instead of submitting an application and hoping for a response, a jobseeker could know what happened next.
That question feels even more relevant to me in a world of AI. If applications become easier to generate and screening becomes more automated, are we helping people connect, or just creating more volume on both sides? Employers still need to understand who's actually qualified. Jobseekers still need a way to get a real conversation and meaningful feedback.
Virtual events solved a real problem and became an important part of the product. But I still believe there's particular value in getting pre-qualified candidates in front of the people who can hire them — in person. For jobs based at a physical location, getting to the workplace was part of what made the original model work.
I think back to those first emails, especially the ones from people who didn't get the job. They still appreciated the experience because they knew where they stood.
That's the part of the vision I still believe in: not just making it easier to apply, but making it easier to get an interview — and to know where you stand.