Swiping Right On InterActive Corp

I am long InterActive Corp in significant size with average cost basis of ~$183/share. In my view, shares of InterActive Corp (IAC US) are a compelling long. IAC is extremely mispriced as Tinder, ANGI HomeServices and Vimeo are substantially under-monetized and have enormous growth runways which are under-appreciated by embedded expectations.

IAC is comprised of its ~81% stake in Match.com (“MTCH”), ~86% stake in ANGI HomeServices (“ANGI”), and its various operating businesses – Vimeo, Applications, and Publishing.

If MTCH monetized at monthly ARPUs of $25 compared to current rates of ~$17, shares would trade at 12x EBIT. While this is a large percentage increase, it would represent a mere increase from 1% of annual dating spend to 1.6%.

If ANGI monetized at 11.5% take rates (vs 3%-4% currently) and 35% adj EBITDA margins (vs ~21% currently), shares would trade at 10x EBITA (amortization is the product of purchase accounting and the result of acquiring durable businesses). These take rate and adj. EBITDA margin assumptions are in-line with marketplace peers, with the former adjusted for supplier gross margins.

If Vimeo earned 20% EBIT margins and was capitalized at low-growth multiples of 10x – 12x, it would be sufficient to justify IAC’s current price (including the above valuations of MTCH and ANGI, and 3x – 5x EBIT on Applications and Publishing, 10x on corporate costs), implying minimal Vimeo growth is being priced into shares.


MTCH is in the business of helping people optimize their sexual strategy through their portfolio of dating brands. Broadly, sexual strategy has two forms – casual and serious. Tinder, PlentyOfFish, and OkCupid, cater to the former whereas Match, Meetic, Hinge, Pairs, and OurTime caters to the latter.

The consensus views MTCH’s portfolio of dating apps as non-durable, largely based on the view success on dating apps results in loss of users – essentially, satisfied (those who form lasting relationships) users are one-time users.

In my view, dating apps ironically results in more unsatisfied users due to numerous psychological biases which many apps seemingly exploit. These psychological biases include the grass is greener syndrome, reciprocation and deprival super-reaction, social proof, and scarcity.

Pre-online dating, one’s dating pool consisted largely of the people in your social circle. This pool expanded when the Internet came along. However, the proliferation of dating apps have resulted in a step-function expansion in one’s dating pool. This has led to the grass is greener syndrome – one’s partner might be perfectly fine but the vastly-expanded dating pool one has access to leads one to believe there is always a better partner.

Unsurprisingly, marriage rates have declined substantially from 1970 to 2016 in every OECD country, average age of first marriages have increased significantly from 1990 to 2016, and divorce rates have mostly increased, according to OECD data.

marriage per 1000

male mean age first marriage

female mean age first marriage

divorce rate

Notably, GlobalWebIndex found that, globally, 34% and 11% of Tinder users were married and in a relationship respectively.

tinder user relationship status

OkCupid data also suggests rapidly increasing social acceptance for casual sexual strategies.

too many partners

too many partners 2

women sexual exploits

women sexual exploits 2

time before sex

sleep before marry

Scarcity, reciprocation and deprival super-reaction tends to be largely a male phenomenon whereas social proof tends to be largely a female phenomenon. This is because males send large numbers of messages whereas females send relatively few. The percentages of first messages which turn into a conversation is far greater for females than males on average. Presumably, getting matches as a female is effortless whereas the opposite is true for males on average.

This is true for all age ranges in the 2009, 2015, and 2017 studies conducted by OkCupid, and is likely true across all dating apps and should continue to be true indefinitely due to the structural abundance of males and scarcity of females from a biological point of view.

messages received vs attractiveness

message reply rates vs attractiveness

percentage first messages turn into conversation

messages sent vs attractiveness

messages sent by age

messages sent vs messages received by gender

first messages sent by gender and orientation

how men and women reply to messages from different ages

Relatively infrequent messaging/matches by females to males invokes a sense of irregular reward, which is a form of reciprocation which triggers deprival super-reaction, as well as scarcity in males. Relatively frequent messaging by males to females invokes validation and social proof.

In my view, the combination of these psychological biases results in a highly sticky user base. Many users tend to go on a ‘Tinder binge’ before deleting the app, only to re-download the app soon enough – rinse and repeat. While this results in high churn, this churn is misleading – when combined with high ‘relationship churn’ by users, this makes for a sticky user base. This user behavior was implied by management on the 2Q16 earnings call. Further, IAC disclosed in its 1Q15 call 50% of their Match US paid users had previously been paid users. I believe this can be largely explained by the psychological biases at play.

Tinder’s paid features essentially self-select for individual user pain-points. Users who are not getting many matches would use Boost; users who travel a lot would use Plus/Gold for Passport; time-sensitive users would use Gold for Picks or Likes You, etc. Because these paid features are ‘personalized’ to solve an individual users’ pain-point(s), it provides enormous value, in my view. Notably, Tinder Gold subscribers accounted more than 60% of Tinder subscribers only roughly a year after its launch (per 3Q18 earnings call) – a strong validation of the value of Tinder’s paid features.

As a result, Tinder has significant latent pricing power and a long runway for subscriber growth, in my opinion. According to GlobalWebIndex data, ~63% of Tinder users (or ~81% if one excludes those who decline to provide such information) are of the mid-50% to top-25% in income, suggesting users are likely to be highly elastic on price. While there is likely some inflation due to response bias, this should be largely offset in the aggregate given the large sample size.

tinder income distribution

MTCH’s consolidated ARPU in 1Q15 was $0.60. This figure has declined to $0.53 over most of the 3Q16-2Q17 quarterly periods, and has recovered slightly to $0.57 in 3Q18. The decline and eventual recovery is not driven by pricing pressure but instead driven by mix-shift from hard paywall brands (e.g. Match) to soft paywall brands (e.g. Tinder). Quarterly average Tinder subscribers have increased from 9% of total subscribers in 2Q15 to 51% in 3Q18.

(Note the decline in non-Tinder subscribers is largely driven by the run-off of Affinity brands (per 2Q17 earnings call) which are uneconomic on an customer lifetime value basis (per 1Q17 earnings call). Non-Tinder average subscribers have largely stabilized in recent quarters, suggesting net non-Tinder subscriber growth ex-Affinity run-off.)

Hard paywall brands tend to have much higher ARPUs as consumer intent to enter a relationship is higher whereas soft paywall brands thrive on word-of-mouth marketing. Hence, hard paywalls by nature (as there is a paywall) require significant paid marketing to drive conversion.

However, soft paywall brands have better overall economics due to substantially lower customer acquisition costs – in its 1Q16 presentation, MTCH disclosed ARPUs of its soft paywall brands were 55% that of hard paywall brands (vs 50% in 9M15, per S-1), but acquisition costs per paid user were 26% that of hard paywall brands (inclusive of respective App Store 30% taxes), and thus soft paywall brands had 89% of the LTV per paid user of hard paywall brands despite ~half the ARPUs.

Consolidated daily ARPU of $0.57 implies ~$17.1/month. Relative to the price of a date (data from Match found the average price of a date in the US is ~$102 and the average single American spent ~$1,596 on dating in 2016, implying ~1.3x dates per month or ~$133 spent on dates) monthly MTCH ARPUs are a relatively small portion (~13%) of dating spending (given significantly lower ARPU, Tinder ARPU is an even smaller portion) but its apps are essential in aiding users in getting a date in the first place. ARPUs are objectively tiny relative to dating spending given ~80% of Tinder subscribers subscribe for 1 month, per the 3Q18 earnings call, making MTCH ARPU ~1% of annual dating spending for such subscribers.

average price of date in US

paytv ARPU

Maslow’s hierarchy of needs suggests friendship/intimacy/family are far more important than entertainment (sub-set of self-actualization). Pricing for PayTV subscriptions are $80-$120/month, though PayTV is generally regarded as a melting ice cube as consumers shift online. While MTCH being able to command similar ARPUs as PayTV is not my base case, the wide disparity between the ARPUs and relative importance between the two indicates significant latent pricing power for MTCH.

As Tinder’s total addressable market (“TAM”) penetration remains low (est. 50m users vs 600m singles ex-China), TAM is potentially underestimated because of the exclusion of non-singles (as noted above, a significant portion of Tinder users are non-singles), and paid user penetration is low (~4.1m average subscribers vs est. 50m users; OkCupid 2015 paid user penetration was 5.7x that of 2012 while MAU doubled and Tinder’s first-year increase in paid user penetration post-monetization was greater than that of OkCupid in 2012, per MTCH S-1), MTCH can likely grow its Tinder’s subscriber base indefinitely.

Replication of MTCH’s core dating brands is difficult due to the need for network effects stemming from initial user liquidity. Capital is not the limiting factor – dating apps with high marketing spend (i.e. hard paywalls) tend to have inferior economics as a result whereas dating apps with the best economics tend to stem from viral adoption which requires little to no marketing spend (i.e. Tinder).

Incremental compliance costs as a result of GDPR (which the company estimates at $10m-$15m per 3Q18 earnings call) and other related privacy regulations have also raised barriers to entry; while these costs are not large, they are daunting from the perspective of a new entrant.

Risks of new dating apps usurping MTCH’s position is mitigated by user compartmentalization of different apps and a largely unsaturated market. Users tend to compartmentalize different dating apps for different sexual strategies. For example, Tinder/Bumble are geared towards casual sexual strategies whereas Match/Coffee Meets Bagel (“CMB”) are geared towards serious sexual strategies. As a result, the average user uses more than 3 different dating apps at one time – per 2Q18 earnings call.

The market for dating apps remains largely unsaturated, with MTCH having 60m+ users as compared to a potentially-underestimated TAM of 600m singles. New dating apps, including Facebook’s initiative, which can convince non-dating app users to try out online dating are thus probably a net positive to MTCH due to app compartmentalization – users trying out CMB may also try out Tinder due to the differing sexual strategies both cater to. As a result, new dating apps likely expand the existing market for all apps by reducing category stigma.

Moreover, MTCH’s scale and willingness to try new ideas (Tinder was incubated by IAC, Chispa cumulative US downloads have outstripped Happn, CMB, and Bumble 1-6 months post-launch, per 2Q18 presentation, etc) makes the company the best-positioned to create the next Tinder, in my view.

Assuming consolidated ARPU of $25/month or ~$0.83 daily ARPU on 3Q18 average subscribers, with the delta relative to 3Q18 consolidated ARPU being solely driven by price at 70% incremental margins due to 30% App Store taxes, shares of MTCH would trade at ~12x EBIT.

While consolidated ARPU of $0.83 daily or $25/month is a large increase from current levels on a percentage basis, it is a small amount on an absolute basis from the perspective of users – as noted above, ~80% of Tinder subscribers are 1-month subscribers, making the cost of a 1-month subscription ~1% of annual dating spending; a $25/month 1-month subscription would amount to ~1.6% of annual dating spending – further price increases beyond $25/month is entirely possible.

MTCH does not expect to be a full US cash taxpayer until 2020.

ANGI HomeServices

ANGI is in the business of helping contractors improve job turnover/time utilization and helping consumers access quality home services at lower prices through its marketplaces. Its moat is derived from network effects between consumers and contractors and economies of scale in sales and marketing. Recreating ANGI requires significant and multi-year investments in a large sales force (to attract high-quality contractors) and marketing (to attract consumers) to initiate the flywheel. Matching a contractor with a consumer is incredibly hard as the marketplace needs to have the exact capacity available at the zip code level. In addition, matching algorithms are improved with experience and data, which favors incumbents like ANGI. Moreover, given ANGI is pricing at 3%-4% take rates (per IAC 1Q17 shareholder letter) which are far below any comparable marketplace business; any new competitor would likely need to sustain operating losses due to the relative difference in scale.

ANGI has a long reinvestment runway. 90% of discovery occurs offline for US home services; consumer awareness is the bottleneck – the company with the largest sales and marketing budget and lowest customer acquisition costs (i.e. ANGI) is highly likely to acquire the most mindshare over the long-term. Per IAC 2Q16 shareholder letter, an average household has 6-8 jobs per year, with ANGI taking ~1.5 jobs, implying significant room to grow its “job-share”.

While skeptics view Amazon HomeServices as potential existential threat, this is overblown given Amazon’s core advantages are in product e-commerce, not service. Even if Amazon were to significantly ramp up its home services business, the market remains heavily under-penetrated, allowing multiple long-term winners. Such a scenario would also likely be a net benefit to ANGI as it would raise consumer awareness and hasten the transition from offline to online. Furthermore, it takes many years to build a large network of high-quality service professionals and attract a large consumer base, as ANGI can attest to.

At the current time, Amazon’s home service business remains nascent. Notably, Amazon charges a 15%-20% take rate on home services, which is severely over-priced relative to ANGI and also gives an idea of potential ANGI take rates at maturity.

In my view, ANGI’s primary competitors are Frontdoor and HomeServe. Frontdoor’s primary market is the US whereas HomeServe is largely UK/EU with some US exposure. Whereas ANGI provides a marketplace connecting consumers and contractors, Frontdoor and HomeServe provide home service plans.

Both Frontdoor and HomeServe are not a threat given ANGI’s TTM sales & marketing spend of ~$540m is roughly twice that of that of both Frontdoor ($255m) and HomeServe (estimated at ~$220m; sales & marketing spend undisclosed, FY17 SG&A is ~35% of OpEx, thus TTM SG&A is ~$280m; sales & marketing spend estimated using Frontdoor sales & marketing expense/SG&A).

This gap will continue to widen given ANGI is growing its top-line at 20%+ (accelerated from mid-teens y/y growth in 1Q18; the sharp deceleration from FY16/17 30%-40% growth rates is due to Angie’s List whose revenues declined ~6% in FY16 per the merger presentation) on a pro forma basis whereas Frontdoor and HomeServe are growing sales at ~10% rates. Management appears to have turned around Angie’s List, which should allow growth rates to reaccelerate to 30%+.

ANGI has substantially lower customer acquisition costs (defined as sales and marketing expense/homeowners or customers) than Frontdoor at ~$31 vs Frontdoor’s of ~$122. While HomeServe’s is ~$31 (~GBP$25) on my estimates of its sales and marketing expense, the company’s growth is much slower than ANGI. Put another way, ANGI has similar customer acquisition costs as a competitor who appears largely focused on maximizing earnings whereas ANGI is maximizing growth; HomeServe customer acquisition costs would likely be much higher if it was maximizing growth. ANGI’s customer acquisition costs is sustainably lower due to its scale and low-cost non-cannibalizing traffic funneling from Angie’s List, in my opinion.

ANGI is severely under-monetized on two fronts as it sacrifices near-term margin to maximize growth and long-term profits. Take rates of 3%-4% are substantially below that of marketplace peers (GrubHub/eBay/OpenTable/Zillow/HomeAway/etc) which generally boast 8%-18% take rates. Further, base rates across multiple verticals (consumer electronics, furniture, auto parts, beauty, shoes, apparel, hotels, and travel) suggest that marketplace business tend to have take rates that are roughly 1/3 of the supplier’s gross margin. Moreover, Frontdoor and HomeServe generate ~$579 and ~$104 in revenue/customer and revenue excl. repair services/customer respectively vs ANGI’s marketplace revenue/homeowner of ~$39 in FY17. Sales and marketing spend is ~63% of sales (TTM, pro forma for Angie’s List acquisition) compared to marketplace peers such as at 15%-45%.

In my view, ANGI’s low take rate and revenue/homeowner can be explained by under-pricing lead generation services and the fact ANGI primarily sells leads (~30% job conversion based on $17b total job value and $3k average job value per IAC 1Q17 shareholder letter) whereas Frontdoor and HomeServe sells jobs. ANGI’s Instant Booking/Instant Connect (~70% job conversion per Macquarie 28Nov18 report) significantly replicates the Frontdoor/HomeServe value proposition to contractors and growing contractor adoption should allow take rates to increase substantially. Per IAC 2Q16 shareholder letter, these features comprise of 10% of volume. Note because ANGI primarily offers lead generation services, its take rates actually decrease when job conversion increases despite increased value to the service professional, resulting in more potential pricing power. Elevated sales and marketing spend is due to ANGI prioritizing growth over near-term profitability, in my opinion.

While Frontdoor is looking to create its own marketplace network, it would be non-competitive vs ANGI given the difference in scale. Moreover, Frontdoor targets lower-risk customers due to its home warranty business and as a result it has a smaller TAM vs ANGI, making its scale structurally capped.

Assuming current take rates are 4% and normalized take rate of 11.5% (roughly 1/3 of home remodeling contractor gross margins per Next Insurance and NAHB; proportion of contractor gross margin is in-line with the base rate of other marketplace verticals as mentioned above and ANGI is largely exposed to remodeling) with 35% normalized EBITDA margins (in-line with LT guidance, half to the entirety of which can be driven by scaling down marketing from ~63% of sales to 15%-45% of marketplace peers, and hence likely conservative) and subtracting estimated share-based compensation and depreciation, shares of ANGI trade at ~10x EBITA.

The large spike in share-based compensation in FY17 is due to the ANGI IPO. Estimated share-based compensation is based on FY16 share-based compensation relative to sales. Amortization is largely driven by acquisition of indefinite-lived assets (Angie’s List) and are hence added back.


Vimeo is in the business of providing content creators a more profitable avenue to monetize their audience. This is done through Vimeo’s cloud-based creator platform which provides content creators with a suite of professional tools in order to create, share, analyse, and monetize their content. These tools appear to be highly valuable to content creators – 50% of Vimeo’s revenue is international, despite international accounting for a fraction of marketing spend per IAC 1Q18 shareholder letter.

Platforms like YouTube can and do get a large portion of the economics of their creators’ content due to their market position. Vimeo provides content creators with better economics – YouTube and Dailymotion takes 50% and 30% of a content creator’s revenue while Vimeo, in addition to a 10% cut, makes money through content creators subscribing to its creator platform.

It allows content creators to monetize their content much more profitably through transactional videos-on-demand (“VODs”) instead of advertisements despite the higher content creation costs of transactional VODs. Per IAC 2Q18 shareholder letter, average revenue per subscriber (“ARPS”) at Vimeo has grown 15% per year for the last 4 years (ARPS was ~$100 per IAC 2Q17 shareholder letter), and is accelerating.

I believe ARPS growth is largely driven by the revenue-share portion (suggesting improving profitability for content creators) and subscribers moving to higher subscription tiers (~30% of revenue is driven by this, per Evercore 3Dec18 report), not the subscription side as I have not seen Vimeo increase monthly pricing for its subscription plans. Content creators on Vimeo tend to experience better economics as they mature, as shown below, per IAC 3Q18 shareholder letter.

vimeo lifetime revenue by cohort

With sufficient scale on the side of the content creator, Vimeo’s creator platform subscription costs are a small portion of a content creator’s revenue but are essential for creators’ revenue generation. This makes Vimeo’s platform an extremely sticky product with highly recurring revenues. Per IAC 1Q18 shareholder letter, Vimeo has 90% annual revenue retention with average subscription life of nearly 5 years. Being a small portion of total revenue also results in pricing power for Vimeo on the subscription side, a lever which remains largely untapped.

While many view YouTube as a competitor to Vimeo, it is actually a supplement to YouTube from the perspective of a content creator, in my view. Because YouTube adopts an advertisement model in order to monetize content, content creators are focused on maximizing views and audience count as they are paid per thousand impressions. This has resulted in many content creators accumulating massive audiences. As Vimeo monetizes content via transactional VOD, it allows content creators to monetize their content more profitably by adding a new revenue stream – assuming $2 revenue per thousand impressions and $6 per VOD, content creators would only need to convert one impression out of 3,000 into a VOD sale in order to earn a similar amount of revenue. Because VODs tend to be so-called premium content in the eyes of the viewer, the conversion rate is likely much higher.

Management expects 20%+ adjusted EBITDA margins in the long-term per Evercore’s 3Dec18 report, which does not appear aggressive given software/SaaS business typically generate similar margins. Given limited capital intensity, EBITDA essentially equals EBIT. Vimeo CEO noted Vimeo recurring revenue run-rate of $160m per Evercore 3Dec18 report. At a 10.0x-12.0x multiple, Vimeo is worth $320m – $384m.

Vimeo is expected to be broken out as a separate segment in 4Q18.

Publishing & Applications

These segments are generally slowly eroding, but are capital-light and extremely cash-generative. While there are some bright spots which could have long-term potential (i.e. DotDash), there is insufficient disclosure to deconstruct said potential, in my view. Hence, I value these segments at 3.0x – 5.0x TTM segment EBIT, implying $190m – $320m valuation for Publishing and $360m – $600m for Applications.

DotDash is expected to be broken out as a separate segment in 4Q18.


Because of the high-growth nature of MTCH and ANGI, instead of valuing these parts outright, I reverse-engineered the assumptions (per above) required to justify their current valuation (MTCH: $42.71/sh, ANGI: $17.10/sh), assuming zero to GDP growth multiples (i.e. 10.0x – 12.0x EBIT multiples, which translate to mid-teens multiples after-tax).

Vimeo, Publishing, and Applications are valued outright, as noted above. $75m in corporate costs per FY18 guidance are capitalized at 10.0x, and IAC’s ~$500m in net cash as of 3Q18 is accounted for. The special dividend for MTCH to be paid in 4Q18 is accounted for through IAC’s stake in MTCH. IAC does not expect to be a full US cash tax payer until 2021, per 4Q17 earnings call.

Per my assumptions, which are, in my opinion, highly conservative, at ~$42/share and ~$17/share MTCH and ANGI are trading at ~12.0x and ~10.0x EBIT and EBITA respectively, and are dramatically mispriced relative to their long-term potential. This view should be expressed through a position in IAC given it trades at a ~10% discount on these assumptions and as it is plausible that MTCH might require financing from IAC in the unlikely event of an exceptionally adverse outcome from the Tinder founders’ lawsuit, which would place IAC in a very favorable negotiating position relative to MTCH shareholders.

Per the proxy, IAC’s CEO Joseph Levin owns ~1.1% of shares and IAC’s chairman Barry Diller owns ~8.2% of shares (most of which are super-voting, thus ~43% voting control). Diller/Levin have a history of spinning off very valuable businesses from IAC (e.g. Expedia, HSN, Ticketmaster, etc) though there is no concrete timeline for a MTCH/ANGI spin. Diller has rights to consent in the event IAC total debt to EBITDA equal/exceeds 4.0x over a 12-month period, which presumably allows him to limit excessive financial leverage. Diller also has a strong track record of shareholder value creation.


3 thoughts on “Swiping Right On InterActive Corp

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s