[Series] 10 Weeks to open a 5-person startup office Paris — Chapter 2 — Recruitment Automation & Process

If you’re looking at opening an office for your startup in Paris, this series will give you every tool, hack & workflow that I used to scale up from nothing to 5 people in 10 weeks. Each section has a bullet-point summary at the top, so feel free to scroll past the story-telling if you just want tips to hire/open an office. I’ll also talk about a few of the key decisions we made and how you can approach those decisions.

If you haven’t already, check out Chapter 1: Spinning up Recruitment to get started.

There are basically three parts to your recruitment workflow: sourcing candidates, reaching out, and dealing with responses (i.e: interviewing). The third part arguably has the most intricacy — interview order, interview question prep, feedback forms & making sure no one gets left — however, the first two parts are where you’re going to spend the most time doing low-leverage tasks: finding candidates, sourcing emails, writing emails, following up, and then getting them into your funnel.

This is me. I’m always talking to candidates ’cause our pipeline is so healthy.

My recruitment stack:

1. Sourcing (outbound): Hiresweet
2. Sourcing (talent platforms): Hired, Talent.io, YBorder
3. Recruiters: Mobiskill, RHezo, Bluecoders
4. Candidate Experience: Lever

While I was spinning up Cold Outreach, I did a sweep of the market for recruiting tools and automation techniques. Our investors Alven Capital introduced us to HireSweet, which ultimately became anywhere from 10–50% of the total outreach we were doing (especially at the end of the recruitment cycle, when we stopped doing cold outreach).

For every open position you have, HireSweet gives you 10–15 candidates each day — depending on the scarcity of talent, my understanding is that they can throttle this number up or down. For each candidate, you get their LinkedIn, StackOverflow, Github, Twitter, Facebook, and their email. HireSweet stores your pre-built outreach template that you honed through cold outreach (one template for each position to make it relevant), and automatically updates first name in the header and a few other things if you want to include them. You can & should validate that each candidate is relevant — if they aren’t, tell HireSweet why with a few checkbox options or custom text and they’ll integrate that into their future propositions. I typically emailed 3–7 of the candidates out of 10–15, which is a highly efficient ration compared to sourcing manually.

Talent Platforms

To compliment HireSweet, we onboarded onto several talent platforms. We had already tried Hired.com in Mountain View, so we were OK committing to paying upfront (Hired has recently switched to a model where you pay to get in and have a quota of positions you can fill), and the other platforms — Talent.io, YBorder, BlueCoders — all operate on a commission-based setup. Hired will send you new candidates every time a relevant candidate takes their profile public, while Talent.io & YBorder still deliver weekly batches of candidates.

For all platforms, understanding their search features is key. For the Data Engineer role, I would regularly search ‘data pipeline’ ‘spark’ ‘scala’ ‘data engineer’ & less relevant terms like ‘Python’ ‘SQL’ ‘ETL’ and ‘big data,’ just to see what came up.

Recruiters

In addition to platforms, I worked with my old friends at Mobiskill — their founders have always been good to me, and if I’m ever going to work with a headhunting firm, it’s going to be with this gang. They have a good reputation among engineers — they fight for candidates and they specialize by technology, so we ended up working with someone focused soley on engineers working with big data technologies. I also worked a little bit with Rhezo, an agency founded by Alexandre Vovan, which has a little less bandwidth but a very good understanding of the engineering ecosystem in Paris. He knew the engineers I loved and he has helped staff the engineering teams I failed to poach, so he’s definitely someone to work with if you’re looking for the market’s top talent.

Before I talk about Lever and the Candidate experience, I should tell you that all three of our data engineer hires came from cold outreach. Feedback from headhunters was that we basically reached out to the entire market (which was true: ~350 candidates contacted), and so many of their candidates (including 2 hires) were already in contact with us when they reached out (or had said no). One of our hires appeared on Talent.io & Hired just days after I did cold outreach and heard back from him (I had to send screenshots once we filled the position to alleviate concerns that we were contacting talent on the side after seeing them on the platform).

Note: the original version of this post stated that we hired all three data engineers from cold outreach. We’ve since discovered that one of our new hires — the one who also appeared on talent platforms shortly after — actually came from HireSweet. We initially missed this due to some tracking issues.

I believe these tools are key for maintaining a solid pipeline: they also influence your sourcing for cold outreach. See a good Data Engineer from AXA? Maybe you should pull all the AXA employees who use Spark. Notice the CTO of a big data startup is suddenly on Talent.io? Maybe things aren’t going so great and you should reach out to the whole team.

Many of our greatest candidates came from unsuspecting places and positions — one of our Data Engineers was the co-founder & CTO of a SaaS startup, as was one of our VP Engineering candidates. As long as you’re willing to say “That’s great that you’re still happy there — we love what you’re doing and we’re always happy to meet and talk about problems around scaling up big data saas companies,” then there’s no real harm in reaching out to currently employed executives (do so, of course, at your own peril, as not everyone takes it the same way).

Here are my feelings on the talent platforms we used for hiring data engineers in Paris, a highly specific role that leverages both generalist technology (Python) and rarely used technology (Scala, Spark, Kafka):

• Hired: I didn’t see a great pool of candidates for the positions we were hiring (VP Eng & Data Engineer. Their platform is by far the most advanced — integration with Lever, very well developer search technology — but at the end of the day, we saw more mobile/front-end/system/devops engineers than data engineers.
• Talent.io: Definitely the best source of interviews for us. They’ve got good coverage of the market.
• YBorder: Pleasantly surprised to find some candidates here are not on other platforms. The product itself still needs a bit of work, but I still believe that there is a need for a platform to connect international talent with the Paris ecosystem, so I’ll continue to use it.
• BlueCoders: Their talent platform is still nascent (they admitted it when I signed up). This is where I spent the least amount of time.

Data Scientists were very popular on all platforms, and I may need to write an article in the future about (1) the difference between the two, and (2) ) why I think that talent that is qualified for data engineering should choose data engineering over data science (hint: there aren’t enough positions for the market in the medium term because everyone is going into it and companies realistically need very few mathematicians, and the good ones start out as engineers).

Once you’ve got your inbound pipeline automated and your cold outreach optimized for relevancy and scale, you’ll quickly need to shift towards focusing 75% of your energy down-funnel. That is, on the Candidate Experience.

I’ll be publishing Chapter 3: The Importance of Candidate Experience tomorrow at 2PM and the rest of the chapters throughout this week.