Description

Elevator pitch

Keyword Studios is the biggest videogame outsourcing company in the world. It is a people business with an average +15% organic growth rate since IPO in 2013 that has been supplemented with a M&A strategy that has 10X EPS since 2014 and has allowed KWS to reach MSD market share of a fragmented outsourced services market.

What started as a house of companies without much cohesion has become a more integrated group that has a presence in every service across the content development lifecycle.

The stock market is a short-term extrapolation machine and what used to be a business trading above 25X earnings due to the attractiveness of a “pick and shovels” play on the videogames industry has now sold off towards 13x my 2024 estimates due to AI concerns and end market slowdown. I believe the later is transient and the former will have a more limited impact that would take years to play out.

While some hiccups are to be expected in the short term, over the long term I expect to see HSD-LDD organic growth supplemented by HSD M&A growth that creates an attractive 15% CAGR compound rate over the 5Y and an investment IRR above those levels with some multiple re-rating towards the high teens.

Business description / deep realities

KWS is an outsourced service provider for the videogame industry that has been built through the rollup of tens of individual studios. They mostly serve AAA clients and have offerings that cover the content lifecycle (from game development to customer support). KWS clients are the biggest publishers in the world such as Ubisoft, Activision, or EPIC. When you think about KWS offerings it is helpful to divide it in 2 buckets:

(1) Game dev, art & marketing – the individual studios that form the group go to market by themselves. Lots of recurring business, reputation, and customer trust (for example Forgotten Empires, a 2022 small studio acquire by KWS that has worked on most of Age of Empires games). This segment does not benefit that much from the central KWS capabilities, but KWS has historically allowed small studios to scale to the next level (see example of Snowed in below).

(2) Postproduction – KWS branded. Work comes to KWS centrally and they distribute it into the right production area based on prices, time zones etc - Centralized BD.

• Cost plus labour model – KWS makes a spread between the cost of labour and the amounts billed to clients allowing for a very healthy 15% margin.

• Management: while the initial owner operator (Andrew Day) that created the rollup engine left in 2021, there has been a professionalization of the business since then. The new CEO comes from Novartis digital division, CFO is ex Flutter, chairman is the former Experian CEO etc. Lack of insider ownership at the CEO/CFO/COO level is a negative tough but we have seen some modest insider buys by CEO over the past weeks.

• Deep realities:

o Global scale allows to win business on the “commoditized” verticals as KWS can better serve clients (in terms of cost and also speed as KWS can “follow the sun”) freelancer networks (testing network), network of audio studios and secure facilities (to prevent content leaks as KWS gets access to games months ahead these are published) in functionality QA authorized by console OEMs

o Reputation is also key as it allows KWS to win repeat business on tasks that seem low value but are actually critical such as localization and testing. AAA budget around +100M and LSD in testing/loc, no one was fired for choosing KWS kind of analogy –

The reason people will tell you that there’s nothing quite like KWS – there are other vendors that do other service lines but no one to the same scale. Someone comes and says ‘we need this solved’ – most of the time KWS can do it. They have strong presence all over the place, they can build out internal teams in a specific office fairly quickly, they can follow the sun if you’re really under the pump” - Tom Slattery – Former Studio Head (Audio – NA).

o On the higher end (game dev, marketing, and some art services that have a higher rev per head) the individual reputation of the studios is what matters as some of the clients don’t even know they are working with KWS. – I would expect the KWS brand to gain more relevance over the coming years.

• Financials

o Normalized organic growth around HSD to + 10% going forward – publisher budgets grow at a HSD rate+ LSD outsourcing tailwinds.

o M&A engine creates 1.7X worth of value for each dollar of capital invested – Buying asset light studios for 10X EBIT4with LDD organic growth. We have seen some multiple inflation as acquisitions have been focused excessively on game development talent but multiples are coming down a bit recently and KWS net cash position will allow them to act countercyclically.

o Net cash position

o Capital light 15% margins business

o Revenue per share has 12X since 2014 and adj EBIT (adjusting for M&A expenses and intangible amort) has 11X since then as organic growth has been closer to 15% and M&A to 20% per anum.

Risks

• AI: other than market short termism this is why I believe there is an opportunity in the stock ATM. The market believes that KWS generates revenue by arbitrating a spread of low-cost labour when the reality is that KWS gets paid to manage complexity on behalf of clients. KWS has not been resting on its laurels and completed 3 acquisitions in the area long before the AI hype began (Kantan, Mighty and Helpshift).

There are several reasons why I believe this is not that big of a concern re terminal value risks:

1. “Only” 40-45 % of KWS 2023 revenues will be potentially disrupted by AI (from my 2023 estimates). On top of that all the incremental M&A is dedicated towards game dev and marketing – I believe game development and marketing will benefit from increased automation and AI such as designers benefited from Adobe or architects from AutoCAD.

These are technical high value-added roles that will be augmented by AI and software automation. I think art is somewhere in the middle as concept art will evolve around these tools (assuming concept art would be commoditized by AI is like thinking that all logo creation would be done by AI and entails a misunderstanding of the creativity aspect of the work)

2. AI could have a significant deflationary impact on some bits of their localization divisions + player support + testing (around 40-50 % of their business). There is a clear distinction in my mind between the activities that publishers want to have and make the game better (art, game dev, marketing) and the activities that you need to have (localization, some bits of testing and player support) where cost is the main priority. The former is inflationary and the later is the opposite. And then there is a clear distinction between the mechanical tasks in the latter bucket and the specialized ones (the mechanical and cheap testing is set to become AI assisted).

Could it be also different from will as I believe this will be a gradual transition that will end up with a business model that combines software and labour and where KWS will play an important role.

3. However even if KWS sees a deflationary impact they are getting paid for the coordination of the activity not for the actual production in Audio and localization testing (they employ freelancers). Sure if this is all automated KWS value added decreases (as there will be less complexity to manage) however I am not convinced that will be the base case and definitely not in the medium term.

4. Localization (KWS automation solution is called Kantan acquired in 2019): AI solutions have existed for years now. AI was useful to translate out of game text. In game is much more complex as it is very context specific and is very sensitive. Moreover localization is a very small percentage of total budget costs. Kantan has some sort of edge now as it can automate translation with a memory of past games, therefore having access to the battery of games is critical (if you have translated FIFA 2019 your engine is much better so it can translate 2020).

5. QA (KWS automated solution is Mighty Games) – There is a bit of a misconception here. The obvious bugs can be automated (Ubisoft is already using AI to automate some testing features inhouse that they were not able to do before) however testing is not that mechanical as the tester sometimes finds new bugs/ gameplay deficiencies that he is not asked to just by playing the game. I believe that the future of testing will be human assisted rather than a total disruption.

6. Customer support (KWS solution is Helpshift): all SAAS is UI on top of databases, and that is what KWS acquired with HS. HS lost the battle vs Zendesk in the horizontal market and KWS acquired the solution to cross sell it to existing customers and offer a more integrated solution. The future of customer support will also be human assisted as some queries can be automated (and were previously automated) and some require agent intervention.

• While I think is very difficult to predict the future and I don’t like being exposed to long term negative optionality I also think that assuming 50% of KWS business will disappear overnight is completely incorrect as (1) most of post-production solutions will be human assisted (2) it will be a long multiyear gradual transition (3) KWS has solutions in the market that according to my conversations industry experts( Lionbridge, former KWS employees ,current QA director at Ubisoft and Universally Speaking CEO) are advanced compared to what is out there.

I believe localization and player support will be the most impacted segments but even in these segments I believe a human assisted solution provided by KWS is a more plausible outcome. Moreover there are tens of small testing shops that will be left behind in this transition and second order thinking leads me to believe that this could be a consolidation opportunity.

• Competition: a few formers have left KWS in the L2Y following Andrew Day departure from the business and now work at competing businesses (Universally Speaking – Andrew Brown, Jaime Giner – Amber Studios etc. The barriers of entry to creating a game development roll up are low but KWS now has a reputed brand and multi country scale.

• Outsourcing: this has been an ongoing trend over the P10Y. The videogame services market has an outsourcing penetration of 30-40%5 however the penetration of postproduction is above 50% vs game dev around high teens. While outsourcing makes sense in both segments (accessing specialized best in class talent / fixed to flexible costs and increasing utilization) the incremental outsourcing penetration could prove to be slower vs the past as most of the low hanging fruit is now gone.

Opportunities

• Collaboration across studios: ATM KWS capacity to do QA does not help them win game development business but I can see a future where the testing /loc starts much earlier and works alongside the game development team through the game creation process instead of working in silos. This is a key priority for the management and would strengthen KWS value proposition.

• Reputation: I see KWS as the opportunity to invest in the leader outsourcing /consulting business of the videogame industry where there is not yet a dominant brand such as in other fields like consulting, IT outsourcing or customer service outsourcing (Accenture, Globant, Teleperformance etc).

Valuation

• 15% revenue CAGR to 2027 – combination of LDD fading to HSD organic growth and HSD M&A fading to MSD (60M revenue acquired over the N4Y). By 2027 game dev, art and marketing would represent 80% of revenue – prudent on both M&A and organic growth.

• Holding EBIT margins at 15%

• If stock trades to DCF value (4% terminal 9.5% discount) multiple should re rate towards 20X which is reasonable if terminal concerns fade away – 20% IRR

• At 15X the stock is a double in 5Y with an IRR around 15%.

How the thesis plays out from here:

It can get worse before it gets better

• Videogame spending is currently a bit flaky – according to Circana digital data is around LSD yoy. While KWS organic growth is more resilient as its exposed to multiyear investment decisions by AAA publishers I don’t rule out organic growth temporarily going towards MSD-HSD (below 10% guidance).

• Slowdowns in organic growth usually come with margin reductions in labour businesses (due to utilization headwinds and companies doing the right thing and keeping talent on the payroll).

• If any of these manifests the stock market will hate it and valuation could close the gap with other “AI impacted outsourcing companies” such as Teleperformance trading at HSD PE.

• These are important points to take into account from a position management perspective. However, playing market timing is a dangerous game so I would rather build the position over time based on expected returns and fundamentals.

• On a longer time frame I expect KWS to go back to 9-10% organic growth rates and start to disclose quantitative data points on its software assisted segments. I also expect to see a higher degree of collaboration between subs and KWS brand to gain more importance in higher end segments.

• The M&A rollup opportunity will continue and I expect KWS to be countercyclical and make use of its balance sheet over these uncertain times. I think we will see an acceleration from the historical 40 acquired revenue per year towards 60-70 (50M YTD).

• I believe KWS will slowly change its value proposition introducing software assisted tools in some segments and becoming closer to end clients in the process (moving from transactional to a collaborative relationship) which could potentially increase margins in the process. I also think AI is overhyped and the actual impacts will be gradual and very different to what the media describes (Google was about to become disrupted by GPT a few months ago kind of narrative).

Catalyst

-