An Elegant Puzzle- Systems of Engineering Management

by Will Larson

  Your reach-out notes can be very straightforward:

  Hi $THEIR_NAME,

  I’m an engineering manager at $COMPANY, and think you would be a great fit for $ROLE (link to your job description).

  Would you be willing to grab a coffee or do a phone call to discuss sometime in the next week?

  Best,

  $YOUR_NAME

  Really, I think it can be that simple! If that seems too simple, then run an A/B test with something more personalized or sophisticated.

  It’s worth noting that you will end up connecting with some folks whom you simply don’t have a great fit for right now. That’s okay. I recommend reviewing their profile pretty rigorously after they accept your connection, and making an honest assessment on fit. If you don’t have something, that’s okay, and it’s a better use of the candidate’s time to not reach out to them. (However, I also think we tend to over-filter on qualities that don’t matter too much! Being respectful of the candidate’s time is, in my opinion, the most important thing to optimize for.)

  Schedule and enjoy your coffees and chats, and remember that even people you aren’t able to work with now are still folks you’re likely to work with next year or next job. Especially in Silicon Valley, it’s a very small network, and you should interact with each person as if they’ll be providing feedback on whether or not to hire you at your next job. (They very well might!) You have two goals for each of these calls or coffees: figure out if there is a good mutual fit between the candidate and the role, and, if there is a good fit, try to get them to move forward with your process. The three things I find most useful to individuals deciding deciding to move forward with our process are describing why I personally am excited about the company and role we’re discussing, explaining how our process works (from our initial chat all the way to them receiving an offer) and then leaving ample time for them to ask questions.

  Keep spending an hour each week adding more connections and following up with folks who have connected. It’s a bit of a grind at times, but it’s definitely a practice that rewards consistency. I’ve found that this is a good activity to do together! Have a weekly meeting of coworkers who come together and source, chatting about how you’re evolving your search. This also helps folks overcome their initial discomfort with cold reach-outs. (It’s worth pointing out that this process is much easier to do with an applicant tracking system (ATS) like Lever6 or Greenhouse,7 which allow you to have a single place to track whether a candidate has already been contacted by someone else at your company. Having a bunch of folks from the same company reach out around the same time can paint a picture of chaos.)

  If you’ve read through this and are quite confident that this approach won’t work, I’m with you: before I tried it, I was similarly certain that it wouldn’t work for me, and that it was just a big waste of time. But I’ve slowly been converted. It’s also important to recognize that it’s very likely this exact approach won’t perform well over time. So try something simple, push through the concerns that block you from starting, and then experiment with different approaches.

  6.3.3 Is this high-leverage work?

  Similarly, a frequent follow-up question is whether sourcing is a high-leverage task for engineering managers. I think it is: candidates are more excited to chat with someone who’d be managing them than they are to chat with a recruiter whom they’ll mostly work with during the interview process. Likewise, I think it’s a valuable signal showing that managers care about hiring enough to invest their personal energy and attention into it.

  That said, I would be cautiously concerned if an engineering manager was spending more than an hour a week on sourcing (not including follow-up chats, as those will take up a bunch more time). There is also a lot of important work to be done closing and evaluating candidates well, in addition to the numerous non-hiring-related opportunities to be helpful.

  As a closing thought, the single clearest indicator of strong recruiting organizations is a close, respectful partnership between the recruiting and engineering functions. Spending some time cold sourcing is a great way to build empathy for the challenges that recruiters face on a daily basis, and it’s also an excellent opportunity to learn from the recruiters you partner with! We’ve been doing weekly cold sourcing meetings as a partnership between our engineering managers and engineering recruiters, and it’s been a great forum for learning, empathy-building, and, of course, hiring.

  Figure 6.6

  The phases of your recruiting pipeline.

  6.4 Hiring funnel

  Most companies believe that they are constrained by funding, product-market fit, or hiring. Books have been written about each of those, and this will be a foray into hiring. In particular, it’ll be a look at how to use the fundamental hiring diagnostic tool: the hiring funnel.

  6.4.1 Funnel fundamentals

  The hiring funnel consists of four major steps: identifying candidates, motivating them to apply, evaluating them for your company, and closing them on joining. Depending on your particular circumstances, any or all of these can be very challenging.

  • Identify

  Candidates tend to come from three large buckets: inbound, sourced, and referrals. Slower-growing and early companies tend to rely heavily on referrals, whereas fast-growing companies tend to exhaust their supply of referrals and to rely more heavily on sourcing and inbound.

  Inbound are candidates who apply to you directly. These come from your jobs website, often administered using an applicant tracking system like Greenhouse or Lever, or from job postings on LinkedIn and other job sites. Inbound tends to be high volume and low quality. The exception is for companies with powerful external brands, typically the culmination of strong product, reputation, and outreach.

  Sourced are candidates whom you proactively find and engage. The most common approaches are using LinkedIn, visiting colleges, and networking at conferences and meetups.

  Referrals are candidates whom someone at your company already knows, typically previous coworkers or friends who attended college together. This tends to be the primary source of hires for smaller companies. At most companies, referrals are the most efficient source of candidates, sporting the highest rate of interviews that lead to job offers.

  • Motivate

  Once you’ve identified candidates you want to consider joining with your company, you need to motivate them to come interview! Some companies prefer to view this phase as a filter to weed out folks who are insufficiently passionate about their work, but I’ve not found that approach very effective. Rather, that approach seems to mostly filter for candidates willing to represent enthusiasm, as opposed to finding authentic passion. Instead, the formula that I’ve found most effective is pretty simple:

  Spend time. Have the people the candidate would work with spend time with them. Grab coffee with them, talk about the projects they’re working on, and get them excited about learning from each other.

  Clearly define the role. Tell the candidate about what they’d be doing, being both very honest and a bit optimistic. That is to say, always give an accurate description of the work, but try to find the best frame for describing the work.

  • Evaluate

  Once you’re blessed with individuals who want to consider working with you, the next phase is to ensure that they’ll be successful additions to your team. This phase is tricky because you’re balancing quite a few objectives, some of which are in conflict:

  Certainty. You want to be as confident as possible that the candidate will be a success in your company. Letting employees go can be hard on morale, and it takes a great deal of time to do well.

  Candidate experience. You want candidates’ motivation to join your company to increase as they’re evaluated, not decrease. One of the worst possible outcomes of your hiring funnel is that you identify folks you want to join your company, but they’re no longer interested in being part of it.

  Efficiency. You also want to minimize the amount of time invested
by both your team and the candidate. How you think about this can lead to significant asymmetries, such as take-home assignments that require significant candidate time but little in-house time for evaluation (well, in principle anyway, since it seems like most folks find take-home assignments quite slow to review thoroughly).

  • Close

  This is similar to the motivate phase, but now instead of asking them to commit a day, you’re asking them to commit a couple years of their life. Many, many factors come in to play, from compensation packages to benefits, to making them feel needed. Because this is the final step, doing well here is an especially important factor in your funnel efficiency.

  When you want to start operationalizing your hiring process, the first step is to craft a process for how candidates will flow through this funnel.

  6.4.2 Instrument and optimize

  Once you have your funnel defined, the second step is to instrument it! This is the most important argument for adopting a formal applicant tracking system, which will provide metadata around your process.

  Instrumentation is so important because it allows you to understand where to focus your efforts. Different companies excel at different aspects, and even the same company will become better and worse at different stages over time. The only way to operate a consistently good hiring funnel is to ground your attention and efforts in your funnel metrics.

  Once you do have the metrics, put your effort where there is the most room to improve. The first step of this is quite literal, looking at the conversion rates across each phase and investing. But what’s slightly less obvious is what the reasonable upper bounds for each section should be. To answer those questions, benchmarking against your peer companies is really the only way to get useful information. If you don’t benchmark, you can find yourself investing beyond diminishing returns in motivating candidates to interview, close, etc.

  Whenever you have a hiring challenge, start from your instrumented hiring funnel and work the problem systematically from there!

  Figure 6.7

  An extended hiring funnel, including onboarding, impact, and retention.

  6.4.3 Extending the funnel

  The funnel we describe above is the most common variant, but I’ve found that a couple small tweaks make it much more powerful.

  Instead of ending your funnel at closing candidates, add four more phases:

  Onboard. How long does it take new hires to get up to speed? It’s pretty tricky to measure this, but since you’re trying to reason about the population rather than individuals, it’s okay to be a bit messy. Pick a productivity metric, maybe commits per week, and see how long it takes for new hires to reach P40 productivity. That’s a sufficiently good measure for understanding how long it takes folks to ramp up.

  Impact. How impactful are the people that you’re hiring? Again, that’s a tricky one to measure, but you’re looking to understand trends, not individuals, so don’t worry about identifying perfect metrics. A reasonably good proxy here is looking at the performance rating distribution for new hires based on time since hiring.

  Promote. How long does it take individuals to get promoted after they’re hired? This is useful for understanding whether employees have access to opportunity8 within your organization.

  Retain. Are the people you hire staying? Fortunately, this one is pretty easy to track by looking at the people who leave. It is, however, quite a lagging indicator, since it typically takes years for folks to leave. Still, I believe that it’s an essential metric to track.

  Not many companies extend their hiring funnel this way, but I think it’s quite useful, reframing hiring from a transactional process into your organization’s lifeblood. It’s also true that many companies are very uncomfortable sharing this kind of information widely. That’s fine. These are indeed quite sensitive numbers, but you should make sure someone is paying attention to them.

  6.5 Performance management systems

  The most sacred responsibilities of management are selecting your company’s role model, identifying who to promote, and deciding who needs to leave. At small companies these decisions tend to be fairly ad hoc, but as companies grow these decisions solidify into a formal performance management system. Many managers try to engage with these systems as little as possible, which is a shame. If you want to shape your company’s culture, inclusion,9 or performance, this is your most valuable entry point.

  The number of approaches to performance management is uncountably vast, but most of them are composed of three elements:

  Career ladders describe the expected evolution an individual will undergo in their job. For example, a software engineer ladder might describe expectations of a software engineer, a software engineer II, a senior software engineer, and a staff engineer.

  Performance designations rate individuals’ performance for a given period against the expectations of their ladder and level.

  Performance cycles occur once, twice, or four times a year, with the goal of assigning consistent performance designations.

  The purpose of these combined systems is to focus the company’s efforts toward activities that help the company succeed. The output of these efforts is to provide explicit feedback to employees on how the company values their work.

  6.5.1 Career ladders

  At the foundation of an effective performance management system are career ladders, which describe the expected behaviors and responsibilities for a role. There is significant overhead to each career ladder that you write and maintain, and also a significant downside to attempting to group different roles onto a shared ladder.

  What I’ve seen work best is to be tolerant of career ladder proliferation—really try to make a ladder for each unique role—but to only invest significant time in refining any given ladder as it becomes applicable for more employees. As a rule of thumb, any ladder with more than 10 individuals should probably be fully fleshed out, but smaller functions can probably survive with something rough. This works particularly well if you extend an open invitation to folks in that role to improve their ladder! (As an aside, I strongly recommend writing a lightweight ladder before you hire the first person into a given role. The alternatives tend to work out poorly.)

  Figure 6.8

  Diagram of a career ladder with multiple levels.

  One effective method for reducing the fixed cost of maintaining ladders is to establish a template and shared themes across every ladder. Not only does this reduce fixed maintenance costs, it also focuses the company on a set of common values.

  Each ladder is composed of levels, which describe how the role evolves in responsibility and complexity as the practitioner becomes more senior. The number of levels appropriate will vary across ladders, function size, and function age. Most companies seem to start with three levels and slowly add levels over time, perhaps adding one every two years. At each level, you’ll want to specify the expectations across each of your values. Crisp level boundaries reduce ambiguity when considering whether to promote an individual across levels.

  Crisp boundaries are also important as they provide those on a ladder a useful mental model of where they are in their journey, who their peers are, and whom they should view as role models. The level definitions are quite effective at defining the behaviors you’ll want in your role models, which are the behaviors you’ll see everywhere a year or two later.

  A good ladder allows individuals to accurately self-access; these ladders are self-contained and short. A bad ladder is ambiguous and requires deep knowledge of precedent to apply correctly. If there is one component of performance management that you’re going to invest into doing well, make it the ladders: everything else builds on this foundation.

  6.5.2 Performance designations

  Once you have the career ladders written, the next step is to start applying them. The most frequent application will be using them as a guide for self-reflection and during career coaching one-on-ones, but you’ll also want to create formal feedback in the
form of a performance designation.

  A performance designation is an explicit statement of how an individual is performing against the expectations of their career ladder at their current level over a particular period of time. Because these designations are explicit, they are a backstop against miscommunication between the company and an employee. However it is a cause of concern and debugging if the explicit designation doesn’t match the implicit signals someone has received.

  Most companies start out using a single scale to represent performance designations, often whole numbers from one to five. Over time, these often move toward the nine-blocker format, a three-by-three grid with one axis representing performance and the other representing trajectory. Having used a number of systems, I prefer to use the simplest representation possible. The extra knobs in more complicated systems support more granularity, but my sense is that they simply create the impression of rigor while remaining equally challenging to implement in a consistent, fair way.

  More important than the scale used for rating is how the ratings are calculated. The typical setup is:

  Self-review is written by the individual receiving the designation. The best formats try to explicitly compare and contrast against their appropriate ladder and level. I’ve also seen good success in the “brag document”10 format.