Impatient Investors Have Created A Screaming Opportunity In AI
The market seems to be misreading the cycle in a major sector. That's created a number of opportunities — with this name the best of the bunch
Highlights:
IT consulting stocks have been crushed of late. A basket of five major names is down more than 35% YTD.
Optimism around generative artificial intelligence seems to be the culprit. Weak results across the sector suggest the trend isn’t helping — and may be hurting.
But there’s a logical explanation for industry weakness: belt-tightening across the corporate world. Simply put, the cycle has turned, and most customers aren’t yet ready for genAI projects to change that.
In the meantime, one sector player trades at what look like trough multiples. A change in the cycle suggests gains well past 100%.
We know that organizations across nearly every industry are working feverishly to understand the impact of generative artificial intelligence on their businesses. We know this because they’re telling us so:
source: Robinhood Snacks1
We also know that most of these organizations have little experience with generative AI. The technology is simply too new. Many companies are now hiring aggressively — and paying well — for workers with experience. But few of those workers have the institutional currency or rank to lead enterprise-wide AI initiatives.
And so there would seem to be an absolutely screaming opportunity for IT consulting firms — among them Perficient PRFT 0.00%↑ — to capitalize on the gap between these organizations’ needs and their in-house talent. Yet that potential opportunity has not materialized, leading the entire sector to underperform (the black line on top is the NASDAQ 100):
source: Koyfin; one-year chart2
Broadly speaking the more tech-focused the consultancy, the worse its stock has performed. That correlation seems surprising, but it actually makes some sense in the context of recent results, which have been awful. Right now, the entire industry seems to be struggling, and the seemingly logical response by investors has been to sell.
But there is an intriguing argument that the market is completely misreading the situation, and failing to realize that AI itself is to some degree the cause of these recent results. Eventually, that will change. GenAI plus a cyclical reversal sets the table for exceptionally strong growth in the not-too-distant future.
If that argument is correct, nearly all of these names have attractive bull cases — but at the moment, Perficient stock seems to have the strongest one.
source: Office Space meme via RobJoswiak.com
Introducing Perficient
Perficient is a project-based digital consultancy. The firm works with clients — some of the bigger names include Ford F 0.00%↑, Caterpillar CAT 0.00%↑, and Johnson & Johnson JNJ 0.00%↑ — on every aspect of their tech operations, ranging from data analysis to digital marketing to internal app development. The focus, unsurprisingly, generally is on verticals outside of tech, where Perficient consultants can add value and expertise to digital efforts, particularly with newer technologies:
source: Perficient investor presentation
The company was founded in 1998, as a provider of what it called virtual professional services organizations. This entailed basically being an outsourcing partner to software companies; within a few years, the business model changed to the current consulting focus. Steady acquisition activity has been a key part of the strategy, with the firm continually buying smaller shops to bring on both talent and clients. Of course, the rapid technological change of the last 25 years has helped enormously as well, creating a nearly-constant stream of new trends through which non-tech businesses need help managing.
Perficient very much was a creation of the dot-com bubble. The company went public the year after it was founded; at the time of the initial public offering it incredibly had only 19 employees. (It would end the year with 180.) Despite that, PRFT was a reasonably good stock over time, providing 8%-plus annualized returns from its first-day close through the end of 2019.
Those returns perhaps don’t sound like much, but of course the July 1999 IPO came less than a year before the tech market peaked. By the end of 2019, PRFT in fact had outperformed the NASDAQ 100, something that few 1999 IPOs can claim (even narrowing the field down to those companies that still exist)3. The optimism, and the outperformance relative to the index, continued through the post-pandemic period, but has reversed sharply of late. PRFT now trades below its level at the end of 2019, with shares down more than 70% from their November 2021 high.
Consultancies Struggle
As the chart above shows, PRFT is not alone in selling off. Across the board, the IT consultancy sector is struggling at the moment.
Accenture ACN 0.00%↑ cut its full-year guidance after fiscal second quarter earnings in March; the firm expects revenue growth of just 1% to 3% in FY24 (ending August). ACN stock is down almost 20% since the release.
The fiscal Q2 report from London-based Endava DAVA 0.00%↑ in late February was even worse. The company’s full-year constant-currency revenue outlook went from 1% to 2.5% growth after Q1 to a 5% to 7% decline three months later. DAVA plunged 42% the next day, is off 53% since earnings and trades at a level not seen since the worst of the pandemic-driven selling in March 2020.
Thoughtworks TWKS 0.00%↑ fell 29% after its Q4 report in February, thanks in large to projections for a second consecutive year with revenue declining double-digits. Forrester Research FORR 0.00%↑ hit a 14-year low this month; it reported a 23% decline in consulting revenue (25% of total revenue) last year. Perficient itself is projecting revenue growth of just 2% to 6% this year on top of a ~flat performance in 2023, and sell-side analysts already have moved to the low end of that range.
The industry right now just looks like a mess. And given that — at least in theory — the group should be benefiting from the huge increase in genAI interest, the negative sentiment toward the space seems logical. After all, if these are the numbers consultancies are posting amid a generational trend, what would business look like without AI demand?
The End Of Consultants
The short-term risk to the group is further earnings weakness. Perficient itself reports Q1 earnings next Monday, May 6, and given an exceptionally ugly chart it’s nerve-wracking to own (or even recommend) shares ahead of that release:
source: Koyfin
Sell-side analysts to some extent are following stock prices down, particularly without a clear catalyst for the sector’s performance to reverse in the next couple of quarters. And so there’s a risk that being long PRFT, or any of its peers, may be correct, but also quite early.
The more credible risk, however, would seem to be that the short-term weakness is a signal from the market of eventual long-term disruption. In a genAI world, consultants may simply go away, replaced by LLM (large language models) that can provide advice of nearly-similar quality at a tiny fraction of the cost.
This is a simple argument to make. But it also appears to be an overly simplistic argument to make, with its weakness highlighted by an analogue from a very different industry.
Back in November 2016, then-President Obama gave an interview to the New Yorker in which he wondered aloud about the effects of technology on societal “cohesion”. Obama pondered a future in which, among other things, “radiologists are losing their jobs to AI.”
As it happened, someone close to me was (and, despite the continuing growth of AI in medicine, remains) a radiologist. I had read the interview, and I asked her if she worried about being displaced. Her response was simple: “No, because if there’s no radiologist there’s no one for the attending physician to blame.”
The same mentality absolutely exists in the corporate world. The idea that a C-suite executive will ask ChatGPT, “how should our company manage our genAI implementation?” is of course fanciful. But even in a world five years from now where the company may have a focused LLM built on internal data, an LLM which may provide advice more specific to the company itself, someone still needs to make the decision.
Consultants have value in the sense of being an outside voice, and of providing trained and (usually) experienced insights. But consultants also provide cover, as Office Space, the movie screenshotted above, so brilliantly demonstrates. In the film, the arrival of consultants is instantly and correctly recognized as signaling a significant workplace reduction; that’s why the consultants are there. In real life, it’s not just layoffs, but process changes, that can be justified as recommended by outside voices.
But of course, the cover can go upward as well. Advice from consultants can provide support for managers as they make recommendations to their higher-ups; that advice means the desired strategy has a basis beyond the opinion of someone not empowered to make the ultimate decision. And, as in the corollary of the radiologist and the attending cited above, consultants provide the potential for face-saving if the plans don’t work.
In all of these cases, the ability to say “well, this is what the consultants think” matters. That ability deflects some responsibility (and in many cases, some personal animus) for potentially unpopular decisions that travel down the organizational chart.
Obviously, consultants do provide some real value; it is helpful to organizations have outside, experienced, and expert voices reviewing processes and aiding in decision-making. But it’s undeniable that the human nature of employees within those organizations also plays a key role in the consistent, large demand for consultants across essentially every industry in existence. To believe that genAI will soon displace consultants requires that human nature be displaced just as quickly. It seems unlikely.
What Is Going On Here?
So if consulting revenues aren’t being pressured by AI, and if the entire sector is down, the obvious question is what has caused Perficient’s revenue to slow so dramatically.
There seems to be a relatively obvious answer. Simply put, the cycle has turned: