Future of Zip Recruiter

Digital Job Board Self-Styled as an “AI Recruiting Platform” Still Subject to Mean Reversion in the Labor Market

 

1) Thesis Description

ZipRecruiter (ZIP) claims to be a two-sided marketplace for job seekers and employers that utilizes a proprietary AI to curate and improve the recruiting process, when, in actuality, the company is more akin to an employer-paid job advertising channel charging SMB’s on a per job posting basis, requiring massive spending in sales and marketing to maintain market share. Even worse, its substantive R&D investment into its AI-based features seems unlikely to offer ZIP any long-term differentiation versus rivals.

The post-pandemic labor market has exhibited over 2x higher unfilled job openings (JOLTS) and almost 2x the level of job turnover (quit rates) than the pre-pandemic average, signaling that many companies were having a hard time finding and retaining new hires. Notably, economists and other financial participants have observed that inflated job openings relative to other labor statistics since the beginning of the pandemic was likely caused by the near-zero cost for employers to post more job listings and likely some amount of labor hoarding to maintain worker supply in a high turnover environment. In turn, the company experienced a surge in the number of employers paying to use its platform (“paid employers”) and in revenue per paid employer (RPPE). However, ZIP has already seen these metrics softening, even as job openings remain elevated at ~55% above pre-pandemic levels with quit rates (a proxy for labor turnover) at ~ 25% above the pre-pandemic averages. That's why many funds have dropped ZIP, but a few top ones still have exposure such as:

Further, the ease of online job posting is likely the cause for elevated JOLTS job openings and assuming high correlation between the metric and ZIP’s revenue may give investor’s false buy signals, while the company’s fundamentals deteriorate from poor strategic positioning and zero-cost alternatives for customers.

ZipRecruiter’s weakness stands out against a backdrop of unemployment near a cycle-low of ~3.7%, which is biased to go higher as the Fed seeks to lower inflation. Rising unemployment would entail employer’s laying off workers with cuts to job postings and HR advertising budgets. However, even with minimal unemployment changes from this point in time, a reduction in labor market turnover would be sufficient to see further weakness in the company’s leading KPIs, driving the stock lower. Meanwhile, consensus still anticipates a ‘hockey stick’ ramp in after ’23.

Additionally, should the labor market go through a downcycle and start to revert to its pre-pandemic mean, ZIP’s value proposition would shrink dramatically, and its recovery would take about as long (12-24 months) as that of the overall labor market, much slower than one would expect from an ‘AI’-enabled digital marketplace. 

The thesis is as follows:

  1. At the current price of $16.57/share, the market assumes 20%-25% revenue decline this year rebounding to 12.5% per annum growth with ~30% EBITDA margins and peak post-pandemic KPIs at the valuation multiple of ~12.5x EV/EBITDA. Valuation ranges from $5.75/share to $23.25/share over five years, with scenarios ranging from a reversion to pre-pandemic mean, all the way to an optimistic ~$1.3B in revenue based on peak KPIs. In our fairly optimistic Base Case, we estimate ~30% downside. Should our highly likely Downside Case unfold, in which the company reverts to its pre-pandemic state, we estimate ~65% downside. We rely on these calculators to arrive at such number:
  2. An opportunity is available due to several reasons:
    1. Company’s market share gains likely capped, as online job recruiting is a largely mature market. Despite implying that the $280B market spent on US job recruiting is available for the taking by the company, online job recruiting only amounts to ~6% market share (~$17B). Moreover, expectations of ~6% per annum growth to ~8% share in 5 years are below original projections of 8%+ per annum growth just a year ago. Notably, market penetration of digital recruiting increased by only 100bps to 6% share during the pandemic. Such a meager response to a very strong demand signal for digital solutions and transactions implies the market ZIP operates in is quite mature and notions of a large TAM capture are unreasonable with larger competitors present.
    2. Massive increase in paid employers on the platform incorrectly interpreted as market share gain instead of the remarkable post-pandemic labor market tightness and surge in job listings. From ’18 to just before the pandemic, ZipRecruiter’s platform averaged ~105K paid employers per quarter, with minimal growth. During that timeframe, unemployment decreased from ~4% to ~3.5%, while jobs openings averaged ~7M (4.4% rate), and quit rates averaged 1.9%. After troughing at ~76K paid employers in 2Q/20, the number of paid employers on ZIP’s platform surged to ~170K by mid-’21 as a strong rebound in demand (and S&M spending) prompted many employers to pay abnormal amounts to accelerate hiring and to use all available means to reach every possible worker.  However, since its direct listing IPO, paid employers on ZIP’s platform have only decreased and are now back down to ~105K as of 1Q/23, while openings are still ~55% above the ~6.5M (3.5% rate) pre-pandemic average.  Notably, inflated job openings relative to other labor statistics since the beginning of the pandemic are likely caused by the near-zero cost for employers to post more job listings and likely some amount of labor hoarding to maintain worker supply in a high turnover environment. This phenomenon has a correlation to the company’s revenue, which is primarily derived from charging per job listing. However, due to ZIP’s higher-cost job postings compared to competitors, a scenario is unfolding where JOLTS job openings remain elevated, and the company’s revenue deteriorates. As an aside, the trajectory of the less-influenced quit rate and the company’s revenue seems much more aligned. In either case, we believe ZIP’s revenues going forward will be structurally lower than many anticipate, as the labor market slows to a more normal cadence containing fewer job listings per employer, with further downside in the event of a downcycle.
    3. Increased revenue per employer seen as improvement in value proposition, however, recent gains are more likely from base-level price increases and employers moving to higher-priced service tiers due to competition for labor. To the company’s credit, while number of paid employers essentially stayed the same from ’18 to early-’20, revenue per paid employer (RPPE) increased ~8.75% per annum from ~$850 to ~1.1K as higher-value enterprise-sized employers grew from less than 10% of revenue to ~20%. It can be argued that ZIP introduced new features employers were willing to pay for, but 75%-80% of its customer base consists of SMB’s, who don’t need sophisticated solutions. Therefore, the likely marginal buyer of these new features were enterprise customers. Throughout most of the pandemic and up until the company’s IPO, enterprise customers stalled out at ~22.5% of revenue and RPPE maintained a ~$1.1K average. Only after the IPO did RPPE increase from $1.1K in 1Q/21 to $1.95K in YE’22 (and even then, enterprise remained at ~22.5% of revenue). This rise was likely due to price increases per job slot, as strong labor demand and wage inflation gave the company some pricing power. Additionally, we believe the increase in RPPE came from greater use of ZIP’s ‘Pro’ plan that charges double the normal job slot rate to essentially implement SEO, putting an employer’s job posting at the top of the funnel on many partnered sites to capture job seekers attention and accelerate hiring times. In effect, an employer’s purchase of higher priced plans is positively correlated to greater competition for labor and turnover. Interestingly, paid employers declined from ~170K to ~105K during this post-IPO period even with job listings increasing for most of that time, implying that the pricing power exerted by ZIP may have caused some paid employers to switch to cheaper alternatives like LinkedIn or Indeed. At the same time, this trend in higher RPPE was probably accelerated as lower-value SMB’s left the platform, coinciding with job openings starting to roll over in late ’22.  Ultimately, then, ZipRecruiter’s RPPE (still almost twice its pre-pandemic level) may be more a function of labor market tightness than value-added improvements. Even worse, ZIP’s attempts at flexing its pricing power generated adverse effects.
    4. ZIP is seen as a ‘disruptor’ instead of as merely a job advertising channel unable to independently assess recruiting efficiency, unlike its larger, cheaper competitors. ZipRecruiter highlights the alignment of its business model with employer success. However, after matching a job seeker with a job posting, it doesn’t have the workflow solutions to know whether that person actually obtained the job. ZIP integrates with many application tracking systems (ATS), but those systems do not have to report results to the company. Further, ZIP’s ‘performance-based’ pricing measure performance not by the number of persons actually hired, but rather by how many applicants click on a job posting.  Without knowing the true outcome of a hiring solution, ZIP can only really measure engagement levels in the search, click and apply steps of the hiring process – and notably, only 10%-15% of job applications clicked-on reach the interview stage. Other business models based on mass outreach and engagement are in essence advertising channels for their customers, not partners. Additionally, we believe the COO’s decade-plus of experience in adtech bolsters our view that ZIP is best thought of as a digital ad channel. By contrast, Indeed Inc, ZIP’s largest competitor, not only has a bigger job board with greater reach and lower prices, but also offers interview/hire workflow solutions. This structure creates an end-to-end solution that can demonstrably improve hiring outcomes and therefore is more attractive to higher-value customers.  
    5. ZipRecruiter’s massive S&M spend indicative of its positioning at the front end of job search process, higher margins from lower spending likely to come at the expense of future market share. Furthering the point above, in order to capture job seekers attention in a highly competitive landscape, ZipRecruiter spends a disproportionate amount of sales and marketing – both on pure advertising to job seekers and employers, and on payments to other sites to display job postings. This shows up in the form of S&M spending amounting to ~65% of revenue on average, from ’19 until the pandemic. Being highly variable, S&M spending was cut to ~40% during the pandemic through to the mid-’21 IPO (to make margins look better), and then ramped back up post IPO to ~60% until mid-’22. To reach enough candidates to achieve its goals, ZIP must continually spend to maintain mindshare among potential job seekers. Importantly, a portion of said spending goes to other job sites to maximize a job’s exposure. Therefore, reductions in spend hinder ZIP’s value proposition as an ad channel, putting it at a structural disadvantage to larger players. In sum, ZIP’s competitive positioning is dependent on and highly correlated with its S&M spending, such that in a downturn ZIP still pays ~40% of its revenue to S&M. Even in a tight labor market, the company has failed to generate even minimal economies of scale, and therefore expectations for structurally higher margins are suspect.    
    6. ZipRecruiter exaggerates its competitive edge in AI.  Competitors with 3x-5x more data should yield superior outcomes. The use of the word AI is almost ubiquitous across the tech industry, and ZipRecruiter highlights its supposed AI abilities more than any other job listing player. However, almost every other job listing and recruiting company is working on ways to use AI to better match employers and job seekers. Except for its chatbot, ‘Phil’, ZIP has little to distinguish itself from competitors when it comes to its AI efforts. Improvement in AI programs from one generation to the next is mostly attributable to the number of parameters they consider and the size of the datasets they’re trained on, and ZIP’s rival Indeed, can train its AI on a resume data set roughly 5 times the size of ZIP’s, is therefore, in a superior position in this area. Moreover, with the advent of ChatGPT there has been an explosion in third-party large language model (LLM) providers that make any legacy chatbot trained on a smaller dataset inferior, unlikely to be a differentiator for job seekers or employers, and a higher cost burden going forward. As such, ZIP is at a dual disadvantage – spending ~15% of revenue on R&D to develop and maintain its AI as larger competitors create better programs utilizing their bigger size, and at the same time LLM’s becoming commoditized as dozens of substitutes hit the market.
    7. Purported shareholder friendly actions such as direct listing in ’21 and share repurchases over the last year, really to let insiders/early investors to exit their positions via a debt-funded company backstop. ZIP went public via a direct listing in May of 2021, with preferred stockholders converting to Class A shares and selling. While a direct listing can be described as shareholder friendly since no new shares are issued, the most likely reason for this method of IPO was that insiders would not be subject to any lockup restrictions. As ZIP’s share price dropped from ~$30 to the low $20s by December 2021, the company conducted an accelerated share repurchase of $50M – and, coincidentally, large institutional investors sold shares as well.  Then in early 2022, ZIP issued ~$550M of high yield notes rated B1 (just below investment grade), ostensibly to invest in its platform and conduct M&A. Subsequently, the company announced a $100M share repurchase in March and an additional $150M in June, followed by $200M in late 2022 and another $100M in early 2023. That’s a total of $550M authorized for buybacks, of which ZIP has already spent ~$500M. During 2022, institutional shareholders continued to sell down their positions, so we believe the buyback was conducted to generate exit liquidity for insiders and maintain a share price in the high teens. Prior to the IPO, insiders owned ~90% of the company in Class B shares (with the remainder owned by employees). They now own ~60% of A/B shares with two PE investors still owning ~25% and looking highly likely to exit their positions in the next year or two. Since its IPO, ZIP has generated ~$235M in free cash flow, with $205M from SBC addbacks. Barring another debt issuance, the company likely cannot generate enough cash to self-fund another share repurchase program of that size. Meanwhile, ZIP is expected to continue to issue SBC on the order of 50%-60% of FCF. We believe any incremental debt issuance at the company is likely a signal that the remaining PE insiders are selling. This level of debt is quite high, especially given how low the average debt level is for Russell 3000's stocks
  3. The largest micro risks in the name on the short side are (a) successful AI and its other products maintaining elevated RPPE, (b) ability to sell its products to job seekers not just paid employers (an actual two-sided marketplace), and (c) its strong market brand allowing for lower levels of S&M spending, resulting in higher margins. The largest macro risks on the short side are (d) labor market tightness, and (e) high levels of job postings and quit rates.

 

2) Business Analysis

A Brief History – Both Destination and Transmission Conduit for Job Postings, Optimizing for Engagement

ZipRecruiter, headquartered in Santa Monica, CA, was founded in ’10 by Ian Siegel, Joe Edmonds, Ward Poulos and Will Redd. The company was built as an online job site, and in ’15 started an R&D department in Israel to develop solutions to address online hiring inefficiencies, namely improving matches between job seeker and employer. Unlike other job boards that are simply destinations, ZIP has the ability to disperse a job posting to a network of partnered job sites to maximize exposure.

The online recruiting process can be segmented into five steps: Search, Click, Apply, Interview, and Hire. The front end of the process includes the first three segments, which require getting a job seeker’s attention with a posting they could be interested in and getting them to create a profile and then apply. Importantly, only 10%-15% of applications reach the interview stage and almost 90% of job seekers never finish an online application. This is the domain in which ZipRecruiter operates.

The interview and hire segments typically require an API integration with an application tracking system that the employer uses. An ATS does not have to send back recruiting efficiency data to ZIP.

The value of ZIP in the front end of the hiring process is to maximize a job posting’s reach on the Internet and to simplify a job seeker’s interaction with it. Where the company has improved upon these basic functions is in the application of ‘AI’ (really just keyword matching) to its database of resumes and job postings. When it finds a good match between a job seeker’s preferences and an employer’s job requirements, it notifies both parties. This system can reduce time to hire and increase the likelihood of a successful hire. However, since ZIP cannot determine the outcome of a match, its value proposition is limited. It does not have the full suite of services necessary to become an actual digital recruiter.  

The company’s revenue is derived from hosting job listings for a fee on either a subscription (70%-75% of revenue) or performance (20%-25% of revenue) basis. Performance fees are typically for enterprise clients and accordingly, SMB comprise 70%-75% of revenue. There are roughly 10.5M jobs on the platform, ~20% remote, and ~80% lower than $60K salary.

Subscription Plans and Pricing

The company’s subscription plans are a succession of one-month contracts with most customers on the platform for a term of 1-3 months. Pricing is based on number of job listings – anywhere from free for a small number of posts, to several thousands of dollars for many. Fees are also dependent on what tier is used: Standard, which simply posts a job to many sites, Premium, which allows employers access to ZIPs database and Pro, which uses SEO to speed hiring for ‘urgent and hard-to-fill’ jobs. Pro is the only plan in which ZIP allows ATS integration. Moving from a Standard or Premium tier to a Pro plan is on average ~2x more expensive.

Performance Plans and Pricing

Performance plans are used primarily by enterprise clients, and pricing is on a cost-per-click model. This is very similar to programmatic ads where the price of the job posting is determined by the demand for screen space and the quality of the data to target a job seeker. Notably, performance has nothing to do with whether someone is actually hired.

 

Management History – Insider Ownership Dropped to ~60% Economic Interest (90%+ Voting Control) Post-IPO, Capital Allocation Strategy Benefits Insiders and Hides Significant Common Shareholder Dilution

Co-Founder and CEO Ian Siegel owns ~14% (~1% ownership via Class A shares), with a ~$1.1M salary of which 50% is a performance-based cash bonus. Mr. Seigel’s previous roles were at Stamps.com and Rent.com as a Web Developer VP from ’98 to ’06, VP of Technology and Chief Product Officer and Pictage.com and MyLife.com respectively, from ’06 to ’11.

Interestingly, MyLife.com, a background information firm, has had numerous accusations concerning false solicitations and inappropriate billing practices. The CEO and Owner, Jeff Tinsley, was sued in 2020 by the DOJ for deceiving consumers.

Timothy Yarbrough was promoted to CFO in December 2021, with former CFO since 2015 David Travers being promoted to President. Mr. Yarbrough has a ~$2.7M salary, of which ~$400K is a cash base and $2M is in stock awards. Prior to being CFO, he was the Chief Business Officer and SVP of Finance at ZIP since 2014. Mr. Travers as President had a similar compensation scheme in 2022, after being awarded $5.5M in stock grants in 2021.

Qasim Saifee was promoted to COO from CMO in late 2021, Previously, he was ZIP’s SVP of Marketing since 2018. Mr. Saifee has over a decade of experience managing programmatic advertiser marketplaces, first as a Senior Director at Yahoo, and then as SVP of Monetization at OpenX for almost nine years.

ZIP’s other co-founders, who are no longer involved in the operations of the company, Joe Edmonds, Ward Poulos, and Will Redd, collectively own ~23% as of mid-2023, down from ~34% at IPO.

The executive team and insiders own ~17.5% economic interest in ZIP (excluding IVP, and Hadley Harbour Masters, a total of ~25%), ~80% are Class B common shares with 20x voting rights.

Concerning capital allocation, the most likely reason for ZIP’s direct listing IPO was to be free of restrictions from insider lockups. Similarly, ZIP’s early 2022 debt issuance layered on high-cost debt not to conduct M&A over time (which would have been strategically beneficial), but rather to provide exit liquidity for insiders. Across the board, insiders sold down their positions post-IPO, to the point where their ownership level went from ~90% to ~55%, including two PE firms who still own ~25%. Given their historical trading patterns, we expect these firms to lower their ownership percentages to the single digits after another year.

Following its $550M debt issuance, the company announced a $100M share repurchase in March 2022 and an additional $150M in June, followed by $200M in late 2022 and another $100M in early 2023. That’s a total of $550M authorized for buybacks, of which ZIP has already spent ~$500M. After increasing shares from ~97M to ~127M from stock awards to management post-IPO, the company has reduced shares outstanding to ~101M, mostly by buying from insiders, offsetting share dilution.

Since the IPO, ZIP has generated ~$235M in free cash flow, of which $205M has come from SBC addbacks (over $25M per quarter or ~10% of revenue). Excluding the IPO share grants in 2Q/21, SBC was ~65% of FCF. Absent another debt-financed buyback program, even a ~30% reduction in the current SBC run-rate would still increase shares outstanding on the order of ~3.85% per annum.

 

Customer Dynamics – Price and Economically Sensitive SMB’s Majority of Revenue, Enterprise Clients Higher-Value and Sticker; Look to Company as Ad Channel to Fill Jobs

Roughly 70%-75% of ZipRecruiter’s revenue is derived from SMBs, many of which are short-cycle customers, posting only for their near-term hiring needs. The remaining 20%-25% of revenue comes from enterprise customers who typically conduct longer, higher-volume hiring programs. Notably, ZIP spends a portion of its S&M budget to attract paid employers, not just job seekers, to its platform.

SMB’s are more economically sensitive than their larger, better-capitalized enterprise peers, leading to greater variability in ZIP’s revenue base and more churn within its paid employer metric. SMB’s typically pay per job slot and don’t require the ATS integration included in a Pro plan. However, with greater demand and turnover in the labor market, these customers are paying up to stay in front of potential workers. ZIP’s Pro plan, which costs roughly twice as much as its standard plan, leans into this urgency in hiring by offering ‘top of the screen’ visibility and other features that send an employer more ‘matched’ job seekers more quickly.  If not for the high quit rate regime currently in place, the better ZIP is at getting an SMB a potential hire, the quicker an employer would leave its platform by dropping the job listing once it’s filled (although the employer would likely return when it needs another new hire).

Larger SMB’s who require ATS integration (and the Pro plan) likely need a person to run the back end of the hiring process – i.e. interviewing and hiring. These are components ZipRecruiter doesn’t have and doesn’t seem intent on developing. As such, even SMB’s using the higher-cost plans use ZIP merely as a means of advertising their job postings. If economic pressures persist or even intensify, they would likely respond quickly by cutting their job ad budgets and possibly number of employees. If ZIP were directly embedded in the paid employer’s HR function, these SMB’s would likely maintain some spending with ZIP even when times are tough. Instead, even facing just moderate economic pressure, paid employers have already left the platform in droves, falling from ~170K in mid-2021 to ~105K in 1Q/23. We suspect most of the leavers were SMB’s.

Enterprise customers (including META) typically require a consistent level of job postings for higher churn positions. However, given the volume of jobs, these customers pay based on cost-per-click (CPC), which is subject to real time supply and demand dynamics like programmatic advertising inventory auctions. Notably, the enterprise clients do not pay for cost per hire; they pay for mere engagement with their job posting. Since ’18, ZipRecruiter has increased its enterprise client base from ~10% of revenue to 20%-25% by ’20, where it has stayed ever since.

While contracts are called subscription, the terms allow for roughly 30 days’ notice on job listings, leading to ZIP having little visibility into their customers plans.

As the labor market either undergoes a traditional downcycle or normalizes (JOLTS openings and quit rates revert to pre-pandemic levels without unemployment stresses), ZIP’s customer base of price sensitive SMB’s would most likely either cut job advertising spending or move down to lower cost plans to save money, while enterprise clients CPC-based pricing comes in simultaneously in real-time. Additionally, the company appears to be challenged to move further up the value chain with customers given their lack of attribution and measurement abilities on its hiring matches beyond simply engagement.

 

Supplier Dynamics – Resume Database and Supply of Job Seekers Critical but Dependent on Sales and Marketing Spend; Minimal Network Effects

ZipRecruiter’s main suppliers are the job seekers who send their resumes for job applications. High numbers of resumes are critical for the company, both to create a database on which to train ZIP’s AI system, and to attempt to establish a network effect that brings employers to the platform and keeps job seekers on the site.

However, job seekers don’t stay on the ZIP platform for long once they have applied for a job (the average is 5-10 weeks). Also, their resumes may become outdated over time, limiting their usefulness in ‘matchmaking’ and AI training. Finally, job seekers may not have directly applied through ZipRecruiter’s website but through a partnered site, which the company pays for. Due to these dynamics, the company doesn’t appear to have substantive network effects and must spend on S&M to bring job seekers back onto its platform the next time they’re searching for a job.

 

Competitor Dynamics – Basic “Search-Click-Apply” Job Boards Have Low Barriers to Entry; Larger Competitors with End-to-End Digital Hiring Solution Add More Value at Enterprise Level

Globally, there are over 50K basic “Search, Click and Apply” job sites, with only a handful directed towards niche job sectors that appear more resilient to new entrants. Most of them instead operate in a highly competitive environment with duplicative postings placed on multiple sites to reach as many potential job seekers as possible. Of the generalist job boards, ZipRecruiter is one of the largest. Its ability to disperse a job posting to affiliate sites is a competitive advantage in that segment of the market.

In the US, the biggest job sites are Indeed (owned by Recruit Holdings, 6098.JP), LinkedIn, ZipRecruiter, CareerBuilder, and Monster. Except for ZIP, each of these other entities has ancillary products and services to keep job seekers engaged on their platform. (In the case of LinkedIn, the network itself is the product, with the job listings as ancillary.)

LinkedIn and Indeed lets employers post small numbers of jobs for free, well below ZipRecruiter’s price point. Moreover, LinkedIn has up-to-date resume information by nature of its social media-like core business and is typically used for finding specialized jobs as ~95% of recruiters use the site. CareerBuilder and Monster are used as a data collection funnel for their parent companies’ ATS and/or other recruiting services.

Notably, Indeed, the largest job site, and the remainder of Recruit’s HR Technology segment, generated ~$7.8B in ’22 and was responsible for ~45% of all online-based hiring in the US in ’20 as opposed to ZIP’s ~4%, according to third party providers. Indeed’s database of over 200M resumes (~50M at ZIP) creates a data advantage that should accelerate AI/ML developments over its smaller competitors. The company is owned by Recruit Holdings (6098.JP), which also owns Glassdoor and several ATS solutions, creating a complete end-to-end solution. Enterprise clients value this simplicity and ability to determine hiring effectiveness.

Recruit believes it can increase its take rate from ~1% of salary to 2%-3% by transitioning its enterprise clients from cost-per-click to cost-per-qualified applicant. With an average salary of ~$60K in the US, an enterprise-level 3% take rate for Recruit for a end-to-end solution equates to ~$1.8K per paid employer, as opposed to ~$1.7K for ZIP’s ad channel.

Importantly, Indeed’s application-to-interview percentage at 4% is over 2x ZIP’s. In an attempt to determine cost-per-hire efficiency with the third-party data (company revenue in millions divided by % market share of candidates hired), ZIP as an ad channel is very expensive at ~$225 per market share percentage hired ($905M in revenue, 4% hired) versus Indeed’s entire HR Technology stack at ~$175 per market share percentage hired (~$7,800M in revenue, ~45% hired).

In sum, ZipRecruiter is better than most job boards for garnering attention, but outside of an urgent hiring environment it’s outmatched on price, scale, AI development, and hiring efficacy by Indeed and a few others who offer cheaper alternatives.

 

Market Trends – Job Posting Market Cyclical, Secular Growth in Online Recruitment Slowing, Appears Mature

Job Market – Openings and Turnover

Since ’80, unemployment has cycled between a near-low of ~3.7% (current reading as of May ’23) to as high as 14.7% (April ’20), averaging ~5.5% through a cycle with durations ranging 4 to 12 years. Job openings per unemployed person are near all-time lows at ~0.6, while the post-GFC era averaged ~3.5. In absolute terms, job openings averaged ~6.5M (3.5% rate) post-GFC, peaking at ~7.5M in ’19 (4.8% rate). Post-pandemic, job openings soared from ~4.7M in early ’20 to over 12M (7.4% rate) in March ’22, which have since declined to ~10.1M (6.1% rate) in April ’23.

Notably, the job opening rate vs unemployment rate (Beveridge Curve) for current levels of unemployment show job openings are ~40% higher than previous cycles (’00-’23 data). This variance in job openings and unemployment is historically anomalous, steering macro economists at the Fed to look at other measures to determine tightness or slack in the labor market (i.e. quit rates). These economists’ postulate that the ease of online job postings has elevated the job openings metric.

Quit rates (as a % of jobs), thought of as labor turnover, averaged 1.9 post-GFC, peaking at 2.4 in late ’19. Post-pandemic, quit rates dropped to 1.5 in early ’20, rebounded up to 3 in early ’22, and have declined to 2.4 as of April ’23, still 30% above post-GFC average. Quit rates can decline without impact on unemployment levels, are correlated to JOLTS job openings and are a proxy for inflationary pressures in the labor market (voluntary job changes typically include a pay increase).

Typical costs to hire an employee averaged ~$4.5K in ’22.

Online Job Posting/Recruiting

As of early ’23, approximately 85% of all job seekers in the US used online resources, almost the same level as in ’15. As a percent of total US HR spending ($280B in ’22 growing ~1% per annum), online job recruiting was ~3% of the market in ’16, increasing to ~5% in ’19. Notably, post-pandemic online job recruiting only amounted to ~6% market share in ’22 (flat from ’20) an increase of 100 bps from ’19, despite rapid adoption in all forms of digital channels. Going forward, online recruiting market growth is expected to slow from ~15% to ~6% per annum and reach ~8% market share in five years.

 

3) Why now?

ZipRecruiter’s business is overearning due to historically anomalous elevated job listing levels and labor turnover in the tight post-pandemic labor market. However, that market, currently under pressure, is predisposed to revert to the mean. Additionally, the company is structurally disadvantaged, having to spend inordinate amounts on sales and marketing to stay relevant to employers and job seekers alike, with minimal pricing power and several cheaper, more effective competitors. With the name trading above 10x EV/EBITDA on elevated expectations of 30%+ adjusted EBITDA margins and the resumption of 10%-15% per annum growth after a few quarters of declining sales, a deterioration in job postings to peak post-GFC levels (from current levels, that would mean a ~20% reduction in job openings and a ~10% in quit rates) would imply modest downside. Should the labor market correct to pre-pandemic mid-cycle fundamentals over an extended period on par with historical norms (meaning a ~30% reduction in job openings and a ~25% in quit rates), ZipRecuiter’s current share price offers material downside risk.

We advocate entering into a short position as the company has already weakened from a decline in job listings and labor market turnover, all of which are still elevated, while consensus remains optimistic on solid adjusted EBITDA generation from SBC and cuts in strategically important S&M spending. Additionally, ZIP’s RPPE (up 15% y/y in 1Q/23) should start to show considerable decline more closely matching declines already seen in paid employers on the platform (down 30% y/y). Further, downside share price pressure should intensify in the coming quarters as ZIP’s debt-supported buyback, which has supported insiders selling their positions, comes to an end and the firm is unable to buoy the share price given minimal FCF generation.

A few key points below illustrate the company’s dubious value potential at this point in time:

  1. Revenue Pressure Even as Post-Pandemic Labor Market Remains Tight and Job Postings Elevated, Normalization Hasn’t Occurred Yet: From ’18 to just before the pandemic, ZipRecruiter’s platform averaged ~105K paid employers per quarter, while unemployment decreased from ~4% to ~3.5%, and job openings averaged ~7M. After troughing at ~76K paid employers in 2Q/20, the number of paid employers on ZIP’s platform surged to ~170K by mid-’21 due to a strong rebound in demand (and S&M spending).  However, since its direct listing IPO, paid employers on ZIP’s platform have only decreased and are now back down to ~105K as of 1Q/23, while unemployment is near cycle lows of 3.7%, and job openings are still ~55% above the ~6.5M pre-pandemic average (and declining). Notably, when job openings proceeded to roll over in 2Q/22, paid employers on the platform followed suit with increasing intensity throughout the year. Even with a tight labor market and elevated job listings, ZIP withdrew its full year revenue guidance after posting a revenue decline of 20% y/y in 1Q/23, and guiding 2Q/23 revenue to a ~30% y/y decline. We anticipate ZipRecruiter’s revenue to decline ~30% in ’23 to ~$640M assuming job openings (and therefore paid employers) decline ~20% y/y to peak post-GFC era levels, optimistically. On a normalized basis (’25-’29), we anticipate ZIP to maintain ~130K employers on its platform, ~10% below its ’21 peak. Should the US labor market normalize with unemployment on a trajectory towards 4.5%-5% and JOLTS job openings ~6.5M, we would anticipate ZIP’s revenue to decline over 35% this year and 25% in 2024 (coinciding with a longer-duration recovery in the labor market), followed by a rebound to 10%+ per annum growth thereafter.
  2. Price- and Economy-Sensitive Customer Base Leaving ZIP’s Platform, Leading Indicator for Higher-Value Enterprise Clients: As detailed above, ZIP’s customer base fluctuated from 105K down to 76K, then up to ~170K within a year and a half following a pandemic lockdown and a heavily supported recovery. However, with inflation, high labor turnover and costs, and rising rates slowing the economy, paid employers are again back down to ~105K.  The volatility of the company’s customer base is due to the fact that 75%-80% of ZIP’s revenue comes from SMB’s, which as a group are highly sensitive to changes in the macro economy and to price changes. Importantly, these customers don’t need the sophisticated staffing technology solutions that larger firms do. As such, with SMB’s rationalizing job listings, we’d anticipate higher-value enterprise clients to exhibit a similar, though moderated, behavior with a lag. Also, since RPPE increases were partly underpinned by a greater percentage of revenue coming from enterprise customers (10% in 2018, 20%-25% ever since), RPPE may decline more aggressively than paid employers if these customers start to rationalize their spending on job listings.
  3. Weak Pricing Power in Best of Times; Peaking Revenue per Paid Employer, a Function of Urgency in Hiring: From ’18 to early-’20, revenue per paid employer increased ~8.75% per annum from ~$850 to ~1.1K as higher-value enterprise-sized employers grew from sub-10% of revenue to ~20%. Only after ZIP went public did RPPE increase from ~$1.1K in 1Q/21 to ~$1.95K in YE2022 (with enterprise customers staying at ~22.5% of revenue). This increase was due to (a) ZIP increasing prices per job slot in the face of strong labor demand and wage inflation, and (b) a greater percentage of paid employers using ZIP’s ‘Pro’ plan. (As we mentioned above, the ‘Pro’ plan charges double the normal job slot rate to implement SEO, putting an employer’s job posting at the top of the funnel on many partnered sites to capture job seekers attention and reduce hiring times.) Given these dynamics, ZIP’s RPPE increases are mostly not from upselling improved software solutions, but from employers forced to pay more to keep a constant supply of job seekers applying in a high-churn labor market. Notably, even with elevated job listings (starting to roll over), low unemployment, and high turnover, paid employers declined from ~170K to ~105K during this post-IPO period, implying that the pricing power exerted by ZIP may have pushed paid employers to switch to cheaper alternatives. In fact, when comparing third-party hiring data, ZipRecruiter is ~30% more expensive as a job advertising site than Indeed, which has both job advertising and hiring solutions. In fact, Indeed’s parent Recruit believes it can ultimately charge 2%-3% of salary as fees (up from 1%) which equates to ~$1.8K per paid employer assuming an average $60K salary in the US. As labor market turnover decreases (using quit rates as a proxy), we anticipate RPPE to stabilize at ~25% off the 2022 peak, or ~$1.5K, which would still be ~35% above the pre-pandemic mean. RPPE declining to pre-pandemic levels of ~$1.1K would seriously jeopardize ZIP’s contention that it's made value-added improvements to its platform.
  4. Normalized Margins Achievable Though Not Durable, Would Come at the Expense of the Future Market Positioning: Prior to its IPO, ZIP generated ~3.5% and ~20% adjusted EBITDA margins in 2019 and 2020, respectively, and issued SBC ~1.5% of revenue per annum. During the pandemic, ZIP was able to generate 30% EBITDA margins by cutting S&M spending from ~65% of revenue to ~40%. In the run-up to the IPO, and in a tight labor market, S&M spending ramped to ~55% of revenue with margins in the teens. Post IPO, SBC issuance increased to ~12.5% and S&M spending scaled to ~52.5% of revenue, only to have adjusted EBITDA margins average ~21% and the number of paid employers on the platform decline. Lower S&M spending likely impacted ZIP’s ability to effectively maintain its price-sensitive customer base, indicating a lack of scale benefits. As such, the company faces a choice between spending less on S&M to maintain margins and losing out on future revenue, or spending more on S&M and watching margins shrink. Additionally, over the five years of available data, ZIP’s margin structure has varied, though it maintained a ~20% average through a cycle as revenue grew almost 2x, implying maturity. So, while ZIP claims it can generate a normalized 30% adjusted EBITDA margin, that ability may be short-lived as lower levels of S&M spending impact future revenue growth, requiring a ‘catch-up’ when coming out of a downcycle. Despite the lack of economies of scale from S&M spending, we believe ZIP can generate adjusted EBITDA margins of 30% (~21% excluding SBC) starting in 2Q/23, albeit with minimal improvement going forward.
  5. Management Fair-to-Middling Quality, Using Company to Cash Out Insiders: While the CEO and co-founders still maintain substantive ownership, the majority of their holdings are Class B shares with 20x voting rights. ZIP’s direct listing with no insider lock-up period, followed by debt-financed share repurchases, leads us to infer that management considers the best use of company capital to be giving preferential treatment to insiders via a clean exit. While insider ownership has declined from ~90% to ~60% post direct listing, two large insiders still own ~25%. Moreover, another debt-financed share repurchase program is possible if insiders push the company to cover their exit in light of deteriorating fundamentals. Notably, the repurchase program obscured the company’s disposition to issue large amounts of SBC, as much as ~4% dilution per annum. Further, SBC addbacks have amounted to ~65% of FCF over the last two years. Barring another debt increase, a share repurchase of similar size would not be possible given low-quality FCF generation, assuming just half is SBC, and share count should begin an upward trajectory once ZIP concludes its current buyback.   
  6. Acquisition Target Possible, Though Minimal Differentiation Even Considering Its ‘AI’: The digital recruiting market amounts to ~6% of total US job recruiting spending, as the hiring process still requires considerable manual input on the back end. So, while the digital market seems mature for job board firms given the meager response to demand signals, the lack of further expansion might be because the problems in moving more spend to digital channels haven’t been solved. This has been the impetus behind larger recruiting companies buying job board firms, such as Recruit buying Indeed in 2012 and Randstad purchasing Monster in 2016. For ZipRecruiter, being at the front end as an advertising channel, the company lacks a complete solution to capture more from the overall market, but a HR/payroll or recruiting company might see value in integrating ZIP’s capabilities into its own suite of services. It is possible that ZIP could purchase a backend HR company which would increase its value proposition; that window of opportunity has likely passed as its debt issuance went to backstop insider as opposed to M&A.  Additionally, despite ZIP’s considerable R&D spending (~15% of revenue) its ‘AI’ solution is neither unique nor superior to those of larger competitors engaged in the same activity. Given these dynamics and precedent transactions (detailed below), an acquisition is unlikely and in the event of one occurring, the clearing price for a slower growing job board firm is at a lower valuation multiple than ZIP trades at today.

 

4) Concerns/Thesis Pressure Points

Employer Behavior Risk

ZipRecruiter’s revenue base is predominantly SMB’s, who are typically lower-value and only require minimal solutions for hiring. Should the company maintain a higher number of SMB paid employers or expand its enterprise revenue percentage, it would signal a change in the underlying value proposition of the firm and its overall stability through a cycle.

Upselling AI and New Product Risk

The company has spent 15%-20% of revenue on R&D per annum to develop its ‘AI’ and new products that facilitate easier interaction between job seeker and employer. While our research indicates that RPPE doubled to ~$1.95K per paid employer due to tight labor conditions, if the RPPE stays elevated through a slackening of the labor market it would indicate that employers are paying increasing amounts for ZIP products.  

New Revenue Source Risk

Should the company’s position be stronger than our research suggests, there is a possibility that ZIP may attempt to charge job seekers for access to its platform, marketing their platform as a quick way to get hired. This would align with management commentary of being a ‘two-sided’ marketplace and would represent a new revenue source outside of our expectations.

Labor Market Risk

Our base case implies a stable labor market with quit rates reverting to norm, which should feed through to lower RPPE. Should the market remain tight, it would improve ZIP’s performance (while should the market go through a downcycle, we would expect both RPPE and paid employers to decline meaningfully). 

Operating Leverage Risk

The company spends over two-thirds of revenue on S&M and R&D, of which S&M spending is ~52.5%. While lower levels of spending preserve margins in downturns, results to date have failed to show scale benefits. However, should ZIP’s ~$1B in S&M spending the last 4 years create brand recognition, the firm could cut spending considerably and improve margins beyond our expectations with no detrimental impact to its market position.

 

5) Business Valuation

ZipRecruiter should exhibit slowing growth this year and a modest rebound after ’24, with metrics like paid employers and revenue per paid employer (RPPE) only 10% off its ’22 peaks. The firm generates revenue through charging employers to post job advertisements on its main site, which can also be transmitted to partnered sites for a higher fee.  

A summary of the Base Case assumptions for the company is below:

  1. Short-Term Guidance: Revenue guidance pulled in 1Q/23 earnings call, maintained $185M adjusted EBITDA in ’23 (mid-point); On a 30% adjusted EBITDA margin for ’23, revenue would equate to $617M (consensus: $696M) if they generate their guided EBITDA.
  2. Long-Term Guidance: Adjusted EBITDA margins of ~30%, Revenue growth ~7.5% per annum.
  3. Total revenue should decline ~30% to ~$640M in ’23 and another 20% in ’24 to $500M per annum and rebound to ~$800M by ’29 with ~130K paid employers on the platform generating ~$1.5K per paid employer.
  4. We anticipate paid employers to bottom at ~88K in 2Q/24 and grow ~7.5% per annum back to ~130K by ’29, 20%-25% above the post-GFC norm and ~10% below ’21 annual average.
  5. We expect RPPE to bottom at ~$1.2K in 1Q/24 and rebound ~3% per annum to ~$1.5K by ’27, ~50% above the post-GFC average and ~15% below the ’22 average.
  6. EBITDA margin of ~28% in ’23 should improve to ~30% ’24 and beyond with SBC at ~10% of revenue, equivalent to a ~30% incremental EBITDA margin. Effective tax rate of 21% in ’23 and beyond.
  7. Capex: ~1.5% of revenue.
  8. ~7.5x normalized EV/EBITDA multiple for ZipRecruiter. Reasoning behind the multiple is in the Peer Analysis section below.
  9. Discount rate at ~15% (small-cap).

 

Five-Year Operating Model

A simple five-year operating model is utilized to determine value.

Base Case:

Base Case assumes a normalization 20%-25% above its post-pandemic peak for paid employers on its system and 50% above peak for RPPE with a ~7.5x EV/EBITDA valuation.

Downside Case:

Downside Case assumes the company reverts to its pre-pandemic norms of ~105K paid employers and RPPE of ~$1.1K at a valuation of ~5x EV/EBITDA.

Upside Case:

Upside Case assumes the company grows to peak post-pandemic levels of ~165K paid employers and RPPE of ~$2.1K on a ~12.5x EV/EBITDA valuation.

 

Peer Analysis – Trading Comps

Zip’s closest public comparables would be DHI Group and Recruit Holdings (6098.JP, Indeed’s parent company), with the former a more accurate benchmark. Additionally, the acquired Monster Worldwide is an applicable comparable and its history as a public company is presented.  The analysis below uses a normalized EV/EBITDA metric to account for how ZipRecruiter should be optimistically valued as its business model reaches a stage of maturity, with lesser variances due to SBC, 10%-12.5% of ZIP’s revenue versus 5%-7.5% of revenue for the group. EV/FCF was not employed given ~55% of ZIP’s FCF consists of SBC addbacks compared to groups ~20% SBC addbacks as a % of FCF.

Public Comparables – EV/EBITDA

  1. DHI Group (DHX) – ~7x normalized EV/EBITDA; currently ~9.5x FTM (’08 trough 3.25x).
  2. Monster Worldwide (MWW) – ~8x normalized EV/EBITDA (’07-’16) (’08 trough 2.75x).
  3. Randstad (RAND.NA) – ~8x normalized EV/EBITDA; currently ~6.75x FTM (’08 trough 5x).
  4. Recruit Holdings (6098.JP) – ~12.5x normalized EV/EBITDA; currently ~14x FTM.

Data for those companies are taken from:

DHI Group (DHX) is an amalgamation of job boards including Dice, specializing in technology jobs, and ClearanceJobs, focused on the niche security clearance job market for government employees and defense contractors. The company has ~9K subscription clients and ~6.5M job seeker resumes while touting its use of machine learning capabilities to optimize job searches. While the company exited several underperforming job boards over the last few years and revenue declined, its core business has improved despite near-term tech-specific pressures.

Monster Worldwide was a large-scale job advertisement platform and resume database with the ability to advertise across social media and through an ad network. After peaking at ~$1.3B in revenue in 2008 the company continued its slide to ~$600M before being acquired by Randstad in 2016. By 2016, Monster had ~51K customers and ~150M resumes.

Randstad is a staffing services company that recruits workers and provides HR solutions for companies. Randstad purchased Monster in 2016 to use its advertisement network and job board to create economies of scale from a portfolio of HR solutions.

Recruit Holdings is a HR technology and business solution company, focused on its HR matching capabilities. The company has digital solutions to address every aspect of hiring and onboarding a worker. Notably, since 2012, Recruit has purchased Indeed (2012), SimplyHired (2016), Glassdoor (2018), and resume.com (2018). Arguably, Recruit is not a true comparable for ZipRecruiter, but it’s indicative of what valuation premium an end-to-end HR firm would command relative to a mere job advertisement company.

The group’s normalized EV/EBITDA multiple equates to ~7.5x and has ranged between 4x-20x from 2008 to 2022. It currently trades at ~10x FTM assuming 0%-5% per annum revenue growth. Zip currently trades at ~10x ’23 EV/EBITDA.

Peer Analysis – EBITDA Margins

EBITDA Margins

  1. DHI Group (DHX) – ~25% normalized EBITDA margin; currently at ~20% FTM.
  2. Monster Worldwide (MWW) – ~15% normalized EBITDA margin.
  3. Randstad (RAND.NA) – ~5% normalized EBITDA margin; currently at ~5% FTM.
  4. Recruit Holdings (6098.JP) – ~20% normalized EBITDA margin; currently at ~15% FTM.

DHI Group’s operating margins have over time decreased from 35%-40% from 2006-2012 to around 20%-25% ever since. Notably, through a cycle DHX has spent 35%-40% of revenue on S&M and ~5% on SBC. Capital expenditures should average 10%-15% of revenue.

Monster’s operating margins remained relatively stable at ~15% post-GFC, never returning to the previous 25%-30% range of 2004-2007. Notably, the company spent 50%-55% of revenue on sales personnel and an additional ~25% of revenue for marketing, a total of ~75% directed towards S&M, but this spending never seemed to yield any scale merits. Capital expenditures averaged ~5% of revenue.

Randstad’s operating margins remained relatively stable at ~5%, lower than the rest of the comparable group given large-scale staffing operations with considerable personnel. Capital expenditures averaged ~0.5% of revenue.

Recruit’s operating margins have expanded from ~10% to ~15% over five years. Notably, the firm spends 10%-15% of revenue on sales and marketing. Capital expenditures averaged ~2.5% of revenue.

ZIP’s comparable group generates ~20% normalized EBITDA margins. ZipRecruiter currently generates ~20% EBITDA margins and in an optimistic view we could see that expanding to ~30% in the near-term, remaining at those levels through a cycle.

Strategic Acquirer Analysis

Precedent Transactions

  1. Monster Worldwide (Randstad in ’16) –~0.75x EV/Revenue, ~6.25x EV/EBITDA.
  2. Careerbuilder.com (Apollo ’17) – ~0.9x EV/Revenue.
  3. Indeed (Recruit Holdings in ’12) – ~5.9x EV/Revenue, ~35x EV/EBITDA; ~30% y/y growth.
  4. Glassdoor (Recruit Holdings in ’18) – ~7x EV/Revenue, ~35% annual growth.

From the transactions presented, strategic assets in a high-growth phase warrant 6x-7x EV/Revenue or 30x-35x EV/EBITDA valuations. Slower-growing, mature assets appear to be valued below 1x EV/Revenue or 6x-7x EV/EBITDA, below where ZIP currently trades.

ZipRecruiter is likely past its early growth stage, and as such warrants a valuation more in line with, if not slightly above, the mature transaction comparables (to be optimistic).

Peer Analysis – Conclusion

For ZipRecruiter, a value investing stock, a ~7.5x EV/EBITDA normalized valuation appears reasonable based on the representative comparable group, with Recruit’s ~12.5x EBITDA an optimistic and unobtainable target given ZIP’s strategically inferior business structure. Notably, the company’s potential strategic value to an acquirer is likely at 6x-7x EBITDA given the firm is more mature and slower growing, slightly below its current valuation.  

 

6) Market Expectations/Perceptions

ZipRecruiter is covered by seven analysts with an average price target of ~$20.50/share, five with Buy ratings and two with a Hold. Investor relations consists of management attending conferences, but actual IR interaction is virtually nonexistent. The third-party handling ZIP’s IR does not answer the phone or have voicemail. The company’s buyback is cited as a benefit for shareholders, an opposing view to our analysis above.

Consensus forecasts flat revenue from 1Q/23 onward with 30%+ EBITDA margins, with growth then inflecting upwards to 15% per annum. Notably, backing into revenue with their maintained $185M EBITDA guide for this year on an optimistic 30% margin equates to ~$617M for ’23, ~10% below consensus.

 

7) Downside Protection – Where’s the Margin of Safety?

The company’s downside is not well-protected even as Zip is trading at modest valuations on optimistic consensus estimates (10x EV/EBITDA). Trough multiples range from 2.75x to 5x EV/EBITDA, or ~60% downside from current valuation. The revenue base is economically sensitive, and ZIP’s relationships with its customers are transactional with short durations. ZIP’s cash flow generation is of low quality as well, with over 50% of FCF due to SBC addbacks (on a backdrop of 3.5%+ annual dilution). The one positive for the company is that in downturns it can cut S&M spending, which is currently ~52.5% of revenue, to protect margins. However, similar business models suggest that the company would be disadvantaged by spending less than 40% of revenue on S&M activities without deeper connections to its customer base.

ZIP might be acquired by a strategic player, although comparables for slower-growing companies like ZIP suggest 5x-10x EV/EBITDA valuations, which does not offer much upside from current levels.

 

8) Conclusion

ZipRecruiter, while marketing itself as an ‘AI’-enabled matchmaker, is in fact similar to other job boards in terms of market maturity and margin structure. The company has benefitted from a historically tight labor market, but this market should revert to its mean naturally (quit rates decreasing) or through a provoked downcycle. ZIP’s ‘AI’ differentiation is minimal, and cheaper, superior competitors are poised to offer superior AI features to the market’s highest-value customers.

We believe a majority of investors are aware of these elements, but place too much emphasis on the qualitative marketing aspects that management highlights, while ignoring ZIP’s questionable overall response to the trajectory of job listings and unemployment.  Further elevated job listings, which can be posted for free at many other sites, may give investors a false buy signal even as the company’s fundamentals deteriorate, providing a compelling entry point to short the name.