Startup Marketing From the Trenches

In this post, I’ll be sharing an update on our marketing efforts at Facet. We’ll talk about our marketing strategy, explain some terminology and give you a view of what it’s like to figure out a marketing strategy at a startup. This post may be a bit more stream of conscious since I’m analyzing our results as I write this.

One of our goals for 2019 is to “Build a marketing engine that produces a steady stream of new clients.” For the past seven years we have relied on word of mouth and personal networking to find new business. That’s one of the main reasons our sales have hit a ceiling - we’ve tapped out that channel and it’s not scalable.

Most founders start by using their personal network to acquire their first customers. Eventually you need to figure out a way to acquire new customers that is predictable and scalable. With the big pivot we made at the end of 2018, we had to find a repeatable way to find new contracting opportunities for the Facet network members.

Our Q1 Plan

Your startup only needs one scalable marketing channel for the first couple of years. Your goal is to find that one channel, and then scale it. Once the one channel is mature, you can then add additional channels.

For us to build a working marketing engine, we just need to find one marketing channel that works for Facet. We started by brainstorming a list of potential marketing channels:

  • Paid Ads
  • Email campaigns
  • Referral Program
  • Speaking at conferences
  • Referrals from existing customers
  • Content Marketing
  • Hunt for contract job openings on job boards
  • Major Account Development
  • Personal Network
  • Meetups
  • Career Fairs
  • Outsourced lead gen companies
  • Developer conferences
  • Social media marketing
  • Influencer marketing

The next step is to test a few of the channels. Each marketing channel you test will take a significant amount of time to test properly, so you’ll want to pick three at most. Here are the three we decided to try in Q1:

  1. Email campaigns
  2. Job boards
  3. Paid ads on LinkedIn, Facebook and Twitter

Email Campaigns

We started by identifying our ideal customer. We thought that engineering managers at software companies from 100 to 5,000 employees would be the most interested in hiring quality software contractors. Using that criteria, we searched LinkedIn for 400 people that matched our ideal customer definition. We used Email Hunter to find email addresses for people we found. Then used Intercom to send out an email campaign to those 400 people. Here were the two emails we sent, spaced four days apart.

Email 1

Subject: Hire ex-Google, Facebook, etc engineers

Hi [first name]​​,

I’m a former Netflix engineer that hates recruiters. So I started a company where engineering managers can easily hire pre-vetted devs, data scientists, and PMs that are looking for contract work. How do we vet? I have high standards, and my first choice when hiring is always to pick someone with several years of experience working at a company that builds software at scale, and has a reputation for engineering excellence. So, we only accept devs if they’ve spent 3+ years at Facebook, Netflix, Google, or a similar company. We help them find quality contract work.

We launched our service for hiring contract engineers several months ago. If you need contract engineers, and you want legit high-quality, we can help you with that right now. We have some of the best engineers in the world actively looking for contract work. For example, Adam and Greg.

You can learn more about Facet at

Let me know if you’d like some help hiring contract devs.

Email 2

Subject: Easy way to hire contract devs that are former Netflix, Facebook, etc FTEs

Hi ​​[first name],

I’m confident that the contract engineers at Facet would meet your FTE hiring bar. They are based in the US (Bay Area, Seattle, LA, etc) and are former Netflix, Amazon, Square, etc devs. They decided to go contract because they wanted more independence. If you need special expertise in something like security, machine learning, devops, you can find it here. Or, if you have areas of your product that you don’t want your core team working on, you can bring in contractors from Facet that will write code you won’t have to throw away. Maybe you just need to hire engineers quickly.

We have something called the Facet Developer Network - it’s a home for FAANGetc devs that have made the switch to contracting. We help them find quality contract work opportunities. You can read more about it in the blog post I wrote about it.

Whether you want on-site or remote, full-time or part-time, there’s probably a dev in our network that will work.

If you are looking to hire contract devs, you should look at some of our devs. You could also just pass me along to your internal recruiter and we’ll work with them. If you aren’t hiring contractors right now, please keep us in mind!

Email campaigns should really have four or five emails in the sequence to increase the chances of getting a response. Sometimes it takes several emails to spark interest. We would have written the 3rd and 4th email in the sequence, if the result of the first two had been better.

Email Campaign Results

Our first email in the sequence was sent to 400 people. It had a 39% open rate. That is WELL above the industry average of 15-28%. Our subject line was good: “Hire ex-Google, Facebook, etc engineers”. The reply rate was 1% - really low given our open rate. That means that either the email body was bad or the subject didn’t match the body very well.

A couple of the engineering managers forwarded the email on to their internal recruiting department. That didn’t go well. I had written in my email that I hated recruiters. That was dumb. We got some pretty mean replies too. It can hurt, but it’s part of sales and you’ll need very thick skin.

From the email campaign we got one lead that was promising, but they didn’t end up signing a contract with Facet. They were mostly interested in hiring full-time.

Email Campaign Takeaways

Email campaigns typically have a response rate of 1-5%. Only a subset of the people that reply will actually be interested. Only a subset of the people that are interested will end up scheduling a phone call with you. If your response rate is 1%, then you might have to email 500 people to get a single sales lead. If you close one out of every ten sales leads, then you have to email 5,000 people to get a single sale. That’s a lot of emails.

My assessment of the Facet campaign is that there is very high interest in hiring ex-FAANG full-time, but not as much interest in hiring contract at the companies we targeted. We need to spend more time figuring out who our ideal customer is. What kind of company wants to hire the best and brightest on a contract basis?

We met with DiscoverOrg to get a bid on what it would cost to buy an email list, since manually compiling a list is so time consuming. It ended up being $15,000-$30,000/year. Since our results were in range of industry averages, email campaigns could be a viable channel for us. Ultimately, the large upfront cost to buy email lists was just too much for us. Dead end for now.

Job Boards

We thought we could look for contract jobs on job boards and fill them with Facet contractors. What we found is that most contract job postings are created by staffing agencies, not the companies with the openings. Staffing agencies generally have a markup of 40%-100% on the rates they charge companies, which is 2-5x the 20% markup Facet charges. If we had direct relationships with the companies, they could hire through Facet and get better engineers for the same price. But subcontracting through another agency just won’t work.

We killed this marketing channel pretty quickly so we could focus on working directly with companies that are hiring. This was a total dead end.

Paid Ads

We tested ads on LinkedIn and Facebook.


We experimented with several ad campaigns on LinkedIn. Before creating our ads we spent some time looking at the ads in our LinkedIn feed to try and figure out what made an ad stand out. I have no idea what makes a good ad, so we just copied ad formats from companies that seemed to know what they were doing.

We created several ads that looked something like this:

Hire ex-Google Ad

We spent $5,000 on LinkedIn ads. For that spend we got 123,000 impressions, 977 clicks and 6 conversions/leads. Our click through rate is .8% - not great, but we are going after a very narrow market, so we think it’s about right. We had trouble setting up the conversion tracking so we might have missed counting some leads that came through LinkedIn. Counting just the 6 signups, LinkedIn says that our cost per lead is about $650. I’m not sure how they calculate that, but if we spent $5,000 and got 6 leads it seems like our cost per lead should be $830.

$830 is a lot of money, but welcome to the world of marketing. It costs us $830 just to have the opportunity to sell someone on Facet. Our cost per click was about $5. We pay LinkedIn $5 for every person that clicks on our ad and visits our site. (Please don’t click on our ads if you aren’t a potential customer or contractor!)

Our click through rate was high on these ads, but I think we made the same mistake here as we made in our email campaigns. Our ad said “Hire ex-Google” in big bold letters and only mentioned contractors in the fine print. We think we got a lot of clicks from people who were looking to hire full-time employees, not contractors.

$5 per click feels like so much money to spend, but it’s all worth it in the end if the cost to acquire a customer is less than the money you can make from them. If the cost per lead is $830, and 1 in 10 leads ends up hiring a contractor for 12 months, then the math works.

MCAC (Marketing Customer Acquisition Costs) = $830 x 5 = $8,300

ACV (Average Contract Value)  = $175 * 40hrs/wk * 47wks/yr * 20% = $65,800

If we could spend $8,300 to generate $65,800 in revenue, then we’d have a marketing channel that works. This is where the old saying of “you have to spend money to make money” comes from. This is also why companies spend so much money on marketing. For example, PagerDuty, which recently filed to go public, generated $80M in revenue in 2018, yet they spent $47M on sales and marketing. Of course, they aren’t profitable yet so take that data point with a grain of salt.

We don’t have enough data yet to know how many leads we need to get a signed contract or what the average contract length is going to be, but I think the numbers above will end up being close to the actual numbers.

CAC should include all costs of acquiring a customer: marketing, referral fees, commissions for your sales reps, sales rep salaries, etc. I’m only looking at the marketing portion of CAC right now. Read this article for a complete guide to calculating CAC.

Facebook Ads

We used the same ad images on Facebook as we did on LinkedIn. We spent a total of $1,600 on Facebook ads. We reached 3,400 people with 54,000 impressions. That means each person saw our ad ~16 times. That generated 126 clicks and 5 client sign ups. That’s $320/lead through Facebook vs $830 for LinkedIn. Facebook seems like the better channel, but I have some doubts about the quality of the Facebook leads. Facebook targeting isn’t as good as LinkedIn.

I’m also not sure if those 5 from Facebook and 6 from LinkedIn were actually 11 unique signups, or if LinkedIn and Facebook both claimed credit for some of the same signups. I need to fix our lead signup to track where the lead came from. With that, I could also see if the leads we are getting from paid ads are actually turning into paying customers.

Paid Ads Takeaways

It looks like paid ads could be a really good channel for us. It’s one thing to see numbers that say spending money on ads works, it’s another thing to dump $10,000/month into advertising. That’s money that could go into our pockets each month. Spending on ads is scary, even when you’ve done testing.

Next Steps With Paid Ads

We’ve created new images for our ad campaigns that we think will make it more clear what Facet does. This should reduce the number of useless clicks from people wanting to hire full-timers.

Facet Revenue as of 2018

Other action items:

  • Implement lead attribution tracking so we know which ad campaign a lead came from.
  • Slowly increase our ad spending.

We’re almost ready to commit to paid ads as our one channel. For now we want to continue testing.

Content Marketing

Content marketing wasn’t on our original list of channels to test, but we sort of stumbled into it.

What is content marketing?

Content marketing is a way to market to potential customers indirectly. Rather than showing a potential customer an advertisement about your company, you provide them with valuable content like tutorials, guides, videos, podcasts, etc. Ever search for how to install Postgres on Ubuntu? You’ll likely end up on DigitalOcean doesn’t sell Postgres or Ubuntu per se, but they do sell a product that is used by the type of person that would need help installing Postgres on Ubuntu. They provide their target customers with a useful tutorial and in return, they get:

  1. Brand awareness: Potential customers learn that there is a company out there called DigitalOcean that sells VPCs and other cloud services.
  2. Credibility: Their tutorials are awesome and it shows that they are experts in the field.
  3. Trust: When the tutorials work, it builds trust in their company.
  4. Customer love: They are helping you for free, which is really nice and creates customers that feel appreciation for them.

Content marketing, if done right, can be a very inexpensive way to acquire customers. You pay the cost to write the tutorial once, and people are viewing it for years to come. DigitalOcean has one of the best content marketing implementations I’ve ever seen.

Viral Content Marketing

In February I published my first real blog post, “Why I turned down my Y Combinator interview”, and it went viral. It drove about 25,000 visitors to the website. We got dozens of contractor signups and several client leads.

It took me about a week to write the post and the results were incredible. Our first instinct was that this would be an awesome marketing channel for us. Ever since that first blog post I’ve tried to create more viral blog posts, but it seems to backfire. It seems like people can tell when you are trying. I don’t know how to predictably create viral content. I don’t think anybody does really…except for maybe Mr Beast.

Conclusion: viral content is AWESOME, but not repeatable for us.

Content Marketing

Rather than trying to create viral content, we are going to focus on useful content. I’ll be writing about four topics on the blog:

  1. Bootstrapping
  2. Venture capital
  3. Startup advice
  4. Our startup journey

We’ll have a separate “Guides” section on the website for advice and help on freelancing, hiring engineers, etc. The hope is that this will help our target audience of freelancers and drive traffic to our website.


You don’t have a viable business until you have a consistent way to acquire customers. It should be one of your first orders of business. Don’t fall in the trap most technical founders do by focusing on the things that make them comfortable - building the product. That’s the easy part. You need to find your first customers using any means possible. Next, you need to figure out a predictable and repeatable way to find customers.

We are close to having that figured out at Facet. Paid ads and content marketing are the two channels we are zeroing in on. Hopefully we will have them fully implemented in Q2.

Subscribe to the Facet blog

If you are interested in modernizing company culture, hiring best practices or just following our journey, you can get the latest blog posts delivered right to your inbox.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.