Best Industries to Specialize In

I'd echo the earlier commenter, choose sector with most idiosyncratic outcomes, volatility and dispersion. Should also be something that you can see yourself developing expertise in, find interesting or are passionate about.

I'd rank them broadly:

Technology Healthcare Consumer Industrials Financials/Real Estate Utilities

I will add a slightly contrarian perspective/ insight: you also want to choose a sector where you will be differentiated and competition will be relatively limited.

Tech is incredibly crowded because it is the 'obvious' choice. The products are highly visible in popular culture and the space is constantly evolving. The downside is that every schmuck thinks they can have a differentiated understanding of FANG (or is willing to try).

Add to that the fact that much of internet/ growth tech requires limited 'deep' analysis and is 'finger in the air' 'secular growth vs. secular decliner' stuff. While a lot of industrials are 'gdp +/- growth' and you focus on the rest of the model, a lot of growth tech is largely about getting the revenue growth rate correct. This kind of directional guesswork is relatively easy to do (though almost impossible to be particularly accurate at), attracting a lot of people that aren't really about that modelling stuff.

What sectors aren't easily crowded by lazy 'generalist' schmucks? Financials, healthcare, and energy - all require some significant domain expertise.

Overall though, I think I would put my vote on industrial. It's the best of both worlds - a big space that has a lot of things you interact with / understand intuitively (cars, trains, planes, houses, furniture). However, it also involves some domain knowledge asnd some serious analysis below the top line (i.e. cost structures, incremental margins and returns, etc).

Someone above also said 'try to be a generalist as long as possible.' This sounds like a good idea, but often just results in you being not quite as smart as the reserve investor across a number of fields. Its harder to pull off, but I might recommend trying to spend time specializing across multiple sectors (i.e. make at least 1 or 2 sector moves in your career)

Just my two c as someone who's worked across several sectors.

My votes are industrials and healthcare. You'll learn to do actual primary research, and learn how to model fixed/variable costs properly. As implausible as it sounds, those are two skills that genuinely differentiate you as an analyst. Healthcare also teaches you how to trade/invest around news flow.

Generalist analysts frequently get burned in these sectors ('industrials look cheap' - end up buying at the peak of the cycle / 'healthcare is simple' - end up buying Valeant). This increases the value of specialists.

Tech is fun, but most tech, esp. the consumer-facing stuff is just guessing TAM , penetration rates, ARPU, etc. Very difficult to build repeatable edge. Let's say you 'estimated' FB's growth correctly in the past - how does that make you a better investor in TWTR or SNAP in the future?

The commentary from below on tech vs. industrials isn't as simple/complicated as people believe. Tech is incredibly varied - semiconductors for example are very similar to industrials in they are cyclical businesses with GDP+- end demand markets and very detailed focus on incremental margins, fixed/variable costs and inventory accounting. Outside of NVDA and AMD most semis are not pie in the sky revenue growth models that are based on some intangible addressable market in the future. Then you have software. Do you know why half of software trades at infinite earnings multiples? It's because they basically represent NPVs of recurring revenue streams and the best companies can reinvest 100% of their proceeds to get additional revenue. Software is basically a series of present value equations based on growth, churn, pricing and incremental/decremental margins from that pricing. Generalists investors thus get destroyed in software if they are buying/shorting based on multiples on their models.

Then look at internet. Amazon is a capex intensive businesses where there is a ton of modeling you can do on the incremental margins and capacity utilization on new fulfillment centers (on the retail side) and data centers (on the AWS side). The best analysts on AMZN were the ones who can understand the unit economics of new data centers and how to think about the flow through of the various business lines to both gross margins and operating income. GOOG has the same dynamic as AMZN. FB and NFLX are a little bit softer but you can definitely do in-depth work on the margin side by looking at the new engineers, cost per new engineer, and think about how the costs can truly grow with revenue based on the limited pool of demand in the valley.

Then on primary research, tech has ton of opportunities. Talk to the software resellers and partners and see how they are doing and what trends they are seeing. Talk to the fabs in Taiwan to gauge end market supply/demand. Talk to the people at AWS reinvent and Strata to figure out what is happening in the database environment.

Tech is a great space and definitely not just about clicks, eyeballs, and buzzwords like AI. Ultimately, the best part about tech is that outside of certain segments like semis, they are predominantly high quality businesses ("compounders" though I hate that phrase as it has become a catch all for everything that is 15x ebitda these days) with long-term growth opportunities. Unlike industrials, you do not need to time a cycle right to do well because over a long enough period of time your stocks typically go up because the intrinsic value of the businesses are growing and the multiples are not high enough where the multiple compression will mask the underlying growth of the businesses over a multi year holding period - as a result you see a ton of the top LT investors have big weightings in tech