Binge-Hiring Engineers: How We Saved More Than $100k In Recruiting Costs

Four weeks ago, we had 6 full-time employees at Cogniac. We’re a team of engineers and entrepreneurs, building an enterprise computer vision platform in a killer office in San Jose. A few weeks ago, something great happened: We had a sudden and overwhelming boost in the number of customers signed-up and waiting to use our products. Our reaction?

Start with this:

And binge-hire until we get this:

Here we go.

Recruiting is hard. It’s time-consuming. Hiring the best people is highly competitive. Recruiting can be expensive. Ultimately as a small company, hiring requires creativity. In the past few weeks, we’ve found that some things work better than others.

First, we hired friends and colleagues

We’ve exclusively used this recruitment strategy for the past 6 months, and we quickly brought onboard 3 more colleagues full-time. Simple enough. The hiring process is much faster when recruiting friends, former coworkers, and when pursuing warm intros (from investors, in this case).

3 new teammates!

Next, we talked with technical recruiters

I think hired recruiters can be a great option — the calculus is: Pay recruiters $money for preserving your precious and scarce time. You’d be amazed at how many recruiters have reached out to us in the past few months, mostly triggered by our LinkedIn profile updates, so getting information from them was simple: Most demanded between $20k — $50k per hire… Scaling our team even 2x would cost us $100k’s. No thanks.

no change…

Enter LinkedIn

Surely you would think that their $3.6b annual revenue bankrolls a decent inbound recruiting product. And you would be partially correct. Here are our results:

We spent $1690 on LinkedIn Ads for these 2 positions, which resulted in 9 in-office interviews and zero offers. Sigh. (Admittedly, we were trigger-happy with in-office interviews from LinkedIn at first re: unqualified candidates. Customer demand knocking on our front door = we need engineers quickly).

But these numbers don’t tell the entire story. LinkedIn’s Recruiter product… well, it sucks. Horrible. Required Context: We were using job posting ads to generate inbound interest and applicants. Case study: A common task might be to target a job posting to current iOS engineers… good luck.

Even after we restrict to ‘skills’ (which I think is based on those annoying spammy endorsements that everybody blindly rage-clicks in order to remove from our news feeds …), about 50% of the applicants were blatant machine learning mistakes.

Few applicants were even remotely qualified for our posted positions.

Targeting is possible only at very low resolution. Targeting skills/endorsements did not work well for us.

Some other problems: Sharing candidates with other team members is not intuitive. Why would you overlay candidate lists as a clickable button on the core LinkedIn product that’s only accessible via an emailed link? Why don’t you allow the person who posted the ad to share ad analytics with multiple teammates? In the recruiter hiring manager tool, why can’t I click on candidates photos or names to view their LinkedIn profiles? (You have to copy/paste their names into google search, “person mcperson linkedin” for example, to query their LinkedIn page for review) And so on…

I don’t believe that LinkedIn Recruitment tools are terrible products— they just failed for what we needed. We wanted to generate inbound interest without manually searching through millions of profiles. I think LinkedIn is probably better at guiding manual searches for very specific niche jobs (i.e. headhunters). For small companies like ours trying to fill a few common but key positions, LinkedIn wasn’t very helpful.

no change…

Finding success with AngelList

I used to think that AngelList only contained a portfolio of buzzwords, but in the past few weeks I have completely changed my mind. AngelList is a brilliant product and useful for companies, candidates, and investors. The ‘news feed’ is spectacular, the search interface for candidates and jobs is tremendous, and the recruitment/hiring funnel tools are easy to use across our organization.

AngelList does offer profile stats, but no job-specific stats yet. Also the recruiting stats seem a bit low in this dashboard.

Here is our AngelList funnel. We extended 2 offers so far, with more forthcoming in our current pipeline.

AngelList’s search-based product is difficult to statistically compare with LinkedIn’s ads product, but the ultimate results are an important metric: We have given 2 offers via AngelList introductions (and a few more forthcoming asap).

Once again, there’s more behind these numbers.

  1. The AngelList product is extremely intuitive, easy to use, and we’ve even seen a couple of great product micro-iterations in the past few weeks.
  2. The quality of technical candidates is tremendous. LinkedIn’s 400M users are diverse — people working all around the world in broad operational, service, and some technical roles. But AngelList is a focused group of 866k mostly tech-focused companies and 485k mostly tech-focused people/candidates. If you’re hiring a fast-food employee in the midwest, don’t use AngelList. If you’re hiring software engineers in the Bay Area, give it a shot.

If you’re looking for a software engineering position at a rapidly-growing company:

Cogniac is still hiring! Check out our website at cogniac.co, see our job postings at https://angel.co/cogniac/jobs, or reach out to me directly at patrick@cogniac.co. You too can claim one of these grey positions:

Cogniac is still hiring!

As we continue to grow our team at Cogniac, we’ll be sure to share more of our experiences. For reference, I have no connections to AngelList or LinkedIn.

About Cogniac

Cogniac enables enterprise customers to extract information from ever-increasing streams of video and image data. Our product uniquely combines the latest artificial intelligence research with human-computer interaction tools and cloud data management to solve visual data science challenges. Our products automate common tasks such as identifying objects, people, actions, and determining normal (safe) vs abnormal (dangerous) objects and patterns. Our founding team has experience building world-class technology and taking it to market at scale — most notably at Ruckus Wireless ($1.2B IPO in 2012, Founder Bill Kish)

www.cogniac.co

Like what you read? Give Patrick Martinchek a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.