This is one thing everyone I've spoken to in the "marketplaces" industry agrees on:

It's really hard to hire operations talent.

It's also my own experience, after having worked for Lyft, Groupon and consulted to a few smaller companies in the space (like Wind, pictured above).

I learned long ago that there are only two priorities for high-growth online businesses that take up 90% of management's time: a) sales (growing gross bookings) and b) recruiting.

In recruiting, there are two kinds of people marketplace companies need: engineering and operations.

It's the need for operations staff that makes marketplace companies distinct from most tech startups, and presents its own challenges.

What is "Operations Talent"?

In a nutshell, operations is where the rubber meets the road.

Many marketplace companies start out with an aspiration of being online-first β€” or even online-only. Long-term, this makes the most sense, economically.

But to get to that point, they need to start with a lot of physical operations. And sometimes regulations creeps in that demands that the physical operations stay.

For example, Lyft and Uber both started out with the premise of being online-only. Drivers could sign up via the website or app. Passengers would download the app and call rides. The engineering teams of such companies thus focused on highly optimised supply/demand matching, pricing algorithms and autonomous vehicles/AI. In theory, there would be no reason to have any physical sites.

Scooter companies also want to be mostly online. Passengers go about picking up their scooters, using their app only. The public sign up to be chargers through the app, picking up scooters and making money on the side for charging them in their homes. Ideally, it's a totally digital ecosystem. If chargers aren't collecting enough: do more campaigns to recruit more chargers, or adjust the payments. If scooters are moving about the city in an unsustainable way, create incentives for one-way rides between zones.

In practice, it's impossible to work online-only, especially in the early days.

The primary reason you can't start out online-only is that the early phase of a company is mostly figuring out what to do, to automate it later. It doesn't make sense to design a perfect ecosystem from the get-go. You deploy a bunch of people and try running out of a series of Google Sheets systems until one sticks, and then you make that into the product.

But moving on from the early stage, there are a number of other reasons you can't stay online-only:

  • Inspections: For Lyft and Uber, in more heavily regulated markets (like California), drivers needed to visit a site for a vehicle inspection, and get official paperwork.
  • Customer service: Hard questions involving a complex trail of paperwork is much easier to solve in person. Sometimes a driver wouldn't speak the language well enough (and online customer service is limited in the languages they offer). Sometimes they just didn't understand their emails, or had lost them, or never received them due to a technology problem. Meeting in person makes it easier to solve these complicated issues.
  • Rentals: Lyft and Uber started renting out vehicles. To rent a vehicle, a driver has to visit a physical site to pick it up.
  • Fleet maintenance: When companies rent out any vehicles β€” whether cars or scooters – they inevitably have to maintain them. This is done either in outsourced garages, or in in-house workshops. Scooter companies like Bird, Lime and wind maintain their own workshops (for now). Peer-to-peer companies like Getaround and Turo send on-demand fleet inspectors (sometimes called "valets") to make sure vehicles are up to spec, operating, equipped with the unlocking equipment.
  • Charging: Scooter companies (Bird, Lime, Wind, Scoot) need to periodically pull in the vehicles and charge them. For the smaller scooters, this is every day. You have to have a whole operation of people bringing the scooters home.
  • Passenger/driver touch-points: Like for example lost-and-found, or some pick-up/drop-off sites, like at airports.

All of these are examples of ground operations that a technology company needs to work out how to do. The problem is, getting people to organise and manage this work β€” and do it well β€” is no easy task.

The Challenges in Hiring for Operations

Hiring for operations talent in startups - a factory
What it actually looks like to operate a scooter company. You thought you were building an app? Think again; you're building a factory.

Hiring operations management talent in marketplace companies is unique because of the combination of skills needed.

If you're in the investment phase, you need to have people who know how to do proper testing, growing a fleet under pressure, and experimenting in a rational way. This means the management of local teams need to know how to conduct tests, track, measure and even begin to scale a new system.

For example, say you're running a scooter company, and you can't figure out why your fleet isn't being deployed before rush hour in the morning. You can do a number of things

  • Pay the chargers/deployers more to deploy by 8am
  • Avoid paying them at all for scooters not deployed by 8am
  • Penalise them disproportionately for scooters not deployed
  • Pay them a flat rate but tell them you'll fire them if they don't deploy by 8am
  • Hire more chargers and pit them against each other

How do you figure out which of these will work? Partly experience, but partly, knowing how to experiment.

A sensible manager will say something like: I'll hire three charging teams under this new structure, and then over a week will see how they perform against others, tracking how they operate, and then if it's successful, I'll try moving some others to this system, and then if that works roll it out to everyone.

A junior manager will say: We'll propose this system to all the teams and see what they say.

That's just an example. There are many other qualities a good operations manager should have.

The Four Requirements of Operations Talent

From my own experience, and also after speaking to a number of GMs/CEOs of these companies, I've boiled it down to:

  • Action-orientation: "I'll do it myself" attitude: an attitude of self-learning ("I don't know, but I'll figure it out!") and a bias towards action (i.e. rather than "strategy").
  • Analytics: Ruthless prioritisation and structured thinking The ability to think logically, prioritise, and to practise highly quantitative management. The kind of person who never reports issues, but rather reports how they prioritised problems and resources they're putting against them to solve them.
  • Clear Communication: Fluent Communication, Up and Down the Chain: Skills in being able to give sensible instructions to multiple teams, and also to report upwards honestly about a situation. They give answers to people below, and convey they've got everything under control to people above.
  • Leadership: In a chaotic environment β€” People skills that can get the most out of a diverse talent pool, ranging from day labourers to analysts.

The supply of people like this is low, managers and recruiters aren't experienced at where to look for them, and don't know how to judge if a candidate is ideal for the job.

Let's analyse these in detail.

Where Companies Currently Find Operations Talent (and Where This Can Go Wrong)

In every region where these companies operate, there is either a shortage of available talent, or a high degree of competition (especially in the Silicon Valley area).

The main places where companies hunt for operations talent, and the ensuing problems, are as follows:

  • Other similar startups: "Got two years experience at Uber? Interested in joining Lime?" The problem is, supply of these people is low. They're usually already employed. If they're not, it's a bad sign. If the company shut down, they're partly to blame, and they won't have developed much experience with winning teams in that time. Finally, having worked at one company means familiarity with the content, but it doesn't mean they'll be able to adapt to new situations.
  • Business school grads: Plentiful, but very typically, extremely abstract thinkers with high expectations. You do get smart people from business schools, but business schools do not cause people to be smart. What an MBA does cause, however, is high expectations of title, salary, and caliber of work done. Thus a predisposition against a lot of the hands-on work that an operations manager needs to do. A Stanford or LBS MBA grad might find it quaint and funny to be washing cars at 2am once or twice to show their "grit"; after three months of stepping in for another overnight worker who quit with no notice, many would think they perhaps made a mistake.
  • Ex-professionals: Also plentiful, with many aspiring to move into tech or startups. People from firms like PwC, Deloitte and the like can be extremely bright and dedicated, but their work will typically have hovered several levels above what an operations manager actually does. You have to screen carefully for grit.
  • Any enthusiastic young person who loves the industry or brand: These exist in abundance! They'll take low pay, harsh conditions and generally do anything. What you need to be careful of is how willing they are to teach themselves, and how much time they're willing to spend learning new things. They may also be impatient when after a few years, they see people with no experience in the industry hired levels above them, after they've put in blood, sweat and tears for years.

The Solution: Better Sourcing, Screening and Interviewing

There are five phases to finding great operations talent, and I'll go over them quickly here, with more detail below.

  1. Sourcing as Lead Generation and Nurturing: Hiring starts with attracting leads. This begins with a compelling job description that speaks to the right people. You then nurture those leads just as you would in any sales company.
  2. Screening via Testing: Screening is best done via testing. It includes two kinds of test, in which you check both a candidate's aptitude and how they think in a structured, quantitative way.
  3. Structured Interviewing: More than just interviewing, this is probing all the areas you need to know about a candidate to make sure they can do everything you might ask of them in the job. It needs a team effort, and needs to be coordinated and structured.
  4. Presentation interviews: This is kind of an acid test to see if you can work with someone. It tests someone's resourcefulness, communication skills and engagement.
  5. Offer: This is where the recruiting team/management negotiates with the candidate.

More detail below.

Attracting leads

Usually, companies put something up on their job page. They'll ask team members to share a role. They'll do some LinkedIn scouring, messaging people to see "if they know anyone".

The surprising part is that this is usually ALL companies will do. All the people who visit the jobs page aren't captured or pixeled for future jobs. Contacts made aren't added to a mailing list (which would be done with their permission, but still should be done). You know how many times I've visited the jobs boards of my favourite target companies to see if there was finally something available? I'd join their mailing list in a heartbeat.

People might say no, or might unsubscribe, but those people were never going to join anyway. That's fine. Meanwhile, anyone who does want to stay on your mailing list is gold.

Screening via Testing

There's a first stage of screening, which is just making sure the candidate meets the bare minimum requirements to work. Once they've passed that, you have to test them.

Testing is done poorly by most companies, if it's done at all. It should be a test whose goal is to quickly weed out candidates who lack the drive to go and learn, to roll their sleeves up and act, and the ability to analyse and interpret data (usually via visualisations and pre-made reports) and make decisions based on its outcomes.

Testing might sound onerous to candidates, but it's such a good acid test that it's impossible to give up.

And to the right candidate, it has a couple of extra purposes that really help them, and help you know they're the right person.

Firstly, it gives them insight into how the company works. When you dump a bunch of company data and analytics on a candidate and ask them to sort through it, it'll give them a ton of insight into what they actually might do day to day. When they show up to the interview, they'll talk about it with enthusiasm!

Secondly β€” and this might be surprising β€” it's fun. It becomes like the analytical equivalent of asking a candidate who wants to work as a pilot to sit in the cockpit of a flight simulator. People really like cracking codes and puzzles and being smarter than the person next to them. If they talk about it with enthusiasm later on (and do well), you know you have a winner.

Structured Interviewing

You've probably interviewed for companies where multiple interviewers people have asked you the same question: "So, why don't you give me an overview of your CV."

Huge waste of time. It maybe would have been useful once (dubious, because I believe this is only good if you're looking for a good story teller), but it's definitely not useful more than once.

There are various reasons interviewers can ask questions that don't help make a decision. Apart from inexperience with interviewing or the whole process being broken (e.g., no CV sent to the interviewer), the main reasons are:

  • No time to prepare because they're crunched with back-to-backs
  • No list of what they should be looking for
  • No idea what other people have asked

Most companies have a haphazard approach to interviewing, asking current employees to interview candidates and assess them on a broad range of skills. The result is usually a loose β€œfit” decision that is rife with bias. Most people are never taught how to interview. And most organizations don't interview in a structured way. By "structured" I mean: most people don't have a list of things to probe in a candidate so they can get a full and well-rounded picture.

Recruiting for operations is unique because it goes far beyond just identifying problems and solving them on paper. You need to be able to always prioritise what to solve, have a path to solving it and be able to communicate process, impacts and risks up the chain of command.

On top of that, the other criteria for a skilled employee do not go away: it has to be someone you want to work with and who people will want to work for.

Given all that, a structured approach is the only way to get through a large number of candidates in a reasonable amount of time.

Presentation interviews

Companies these days like to give candidates a homework assignment, so they can prepare something and bring it in. These interviews gives the current team members the opportunity to see what it'd be like to work with the candidate on a real problem.

The presentation interview simultaneously tests interpersonal and communication skills and structured thinking. It’s an acid test for whether someone is competent for a job.

Given it's an acid test and there's no pass/fail for them, most companies do a reasonable job of this already.

The one thing to call out is that companies should never make exceptions. For example, they shouldn't say "Well his presentation was a dog's breakfast, but he was charming and knows his stuff." I've seen cases where this was argued in senior hires, and it turned out the work they did was as effective as their presentation was. Worse, their "charm" meant that they ran an organisation based on charisma, rather than on quality work.


The final phase is the offer (and acceptance). I'll leave this to the recruiting team.

What's next?

If you want more info on the above, I'm happy to connect with you and share with you any of my guides on hiring or interviewing for operations. Sign up to my list (I only send stuff like this) and respond to the email to let me know what you want.