When we first had the idea for building a job board designed for developers the market was fairly open, and our biggest competitor was Joel Spolsky's job board which was only a side-business to his blog. Since then things have changed with Joel along with Jeff Atwood founding Stack Overflow with Joel's job board being merged with Stack Overflow to become Stack Overflow Careers (now called Careers 2.0).
No longer was our main competitor a relatively small minnow. StackOverflow have gone on to raise $18 million dollars in investment (with reportedly a nine-figure valuation) and their developer Q&A site has 17 million monthly uniques, Careers 2.0 is their primary income source. We'd become David to StackOverflow's Goliath.
We knew from the start we couldn't compete on huge ad campaigns or massive number of eyeballs, we couldn't compete on being huge, so we decided to compete on being lean. We decide to start by focusing on the UK market and getting intimate with our customers, knowing that what we learnt from them would allow us to achieve organic growth and build a much stronger product than StackOverflow's waterfall approach to business model development.
With a budget of only £250/month + the income CoderStack was bringing in, over four months we've built CoderStack to the point where it's more successful than StackOverflow Careers on most metrics in the UK. We run more job ads, our jobs get viewed by more developers, we get more applications/job and we have more satisfied customers. We're even doing better in SEO terms.
This is the story of how we're doing it
Thinking lean
Too often people think about "lean startup" as a methodology, follow a list of steps and things will work, it's not like that. Being lean is a way of thinking, it has to apply to everything you do. Lean is the application of the agile process to business.
In agile development you write a test case before you write the code that passes it, the same applies with lean. Before we make any change to our business (whether it's adding a feature to our site or buying ad space) we decide how we're going to measure if it's successful.
We don't just build a Minimum Viable Product we build Minimum Viable Features. We started with the most stripped down job board possible, initially you couldn't even search by language or location. We went to developers and asked them what they wanted, with every feature we introduced we started with a cut down basic version to see if it was just something developers wanted or something developers would actually use.
We don't have lots of spare development cycles, so we need every developer hour to count.
There's a 80:20 rule in software development. That 80% of the functionality will take 20% of the work, and the remaining 20% (often dealing with fiddly edge cases) will take 80%. Once we have 80% (often even less) of a feature working we push it out live and see how users are using it. On more than one occasion we made it work with huge hacks. Often we found the users just weren't using a feature so we could scrap it and move onto something else glad we hadn't wasted our time on making it perfect.
An example of this was search. Lots of people we spoke to looked at our site and said "why isn't there a search box", from using other job sites developers had become acclimatized to typing in what they wanted into a box. So we hacked together a quick search using SQL likes (if you're not a developer: this is an incredibly easy to implement way to do search; processor usage however is horrendous and heavy usage will kill your database server under load), we pushed it out and started measuring. It turns out only a tiny fraction of users used search, most were happy navigating via links, we did however find out that the search data on which search terms developers were using provided us valuable insight into where our navigation links were failing.
The value in search wasn't primarily to our users but to us in terms of analytics, so rather than building a sophisticated search engine, we decided to buy a search-as-a-service product from Google that was trivial to plug into our site and gave us great integration into Google Analytics.
Thinking lean means being iterative, we haven't done a big launch, we've taken slow but steady growth. In contrast StackOverflow Careers have had two major "big bang" launches, the latest even winning them the "Best Launch 2.0 Overall Company" at the Launch conference (we're not kidding).
The risk with big launches is that they stop you continually innovating and testing as you go along. If you bottle up what you're working on, you're at a much higher risk of failing when you do launch. Imagine if you were developing a piece of software and never ran it to test it worked before releasing it. That's what companies are doing from a business viewpoint when they have big-bang launches.
Customer Development
We learn our customer's pain points and use them to position ourself in the market
StackOverflow Careers understand the first point, the primary pain point of developers is having to deal with recruiters (both they and we bar recruiters from using our respective services), the primary pain point of companies is hiring good developers (although companies often have big pains with recruiters too). What StackOverflow don't do is use this information to position themselves in the market, the right positioning can have a disproportionate impact on your marketing effort.
When selling you need to hit your customers on both the logical and emotional level, and it's the later part that Stack Overflow fails on, when they sell their job board, they're not selling to the customers pain point.
We pitch both developers and companies with an anti-recruiter pitch, as soon as we say that we are an anti-recruiter job board we get asked for business cards. Recruiters drive developers and the individuals responsible for hiring crazy, it hard to describe unless you've experienced the frustration of being called repeatedly by useless but desperate recruiters. We've spoken to investors who just can't grok the concept of people hating recruiters, it's one of the reasons we haven't taken outside funding so far.
Not only does being "anti-recruiter" get developers and employers over to our side, it means that the recruiters end up providing us free advertising. The next time the employer or developer is frustrated by an annoying recruiter, they're going think of CoderStack. By clever positioning we're getting our competitors (the recruiters) to act as our promotional machine.
We constantly react to customers feedback about our product and their feedback about our competitors
We speak to a lot of customers on a regular basis, most of our sales are still hand generated by old-fashion sales techniques, we know this won't scale in the long term but what we're learning from it certainly will. We don't just talk to potential customers about what they want from us, but also what they like and don't like about our competitors, and naturally this includes Careers 2.0 and this feedback feeds into our product.
One of the biggest complaints we hear about our competitors is that companies are unhappy about not getting value for money.
Like many other job boards StackOverflow Careers makes the mistake of offering a standardized price across all locations. Without the same value being delivered across all location it's a recipe for frustrated customers. It makes little sense to charge someone in San Francisco the same price as someone in London, if you have ten times as many developers in the first city over the second.
We had to think long and hard about how we would do pricing, we didn't want to have the same problem, but we didn't want the complexity of dynamically pricing every location and skill. We actually ended up modifying our business model to make it work. Companies aren't unhappy about the amount they're spending for ads on job boards, they are unhappy about not getting value for money. So we decided to increase the value for money of the service we provide.
Now if we get a job post in a region where we don't have the traffic to justify the price we're charging we start adjusting our ad campaigns to target that area until we get enough traffic that we can justify that price. About a third of our traffic is from paid sources such as Facebook, Plentyoffish and Google who all offer geotargeting, so doing this is relatively trivial for us, but means our customers are much happier with our service.
Customer development has to feed into your business model and not just into your product.
We balance the needs of developers with the needs of companies
A job board is a two side market, and in any two sided market you need to be sure you're doing customer development on both sides.
The classic example of single sided customer development is the story of the "dating site for guys", where guys post profiles and women message them competing with each other for who can get a date with the guy. The obvious problem is that no women would want to use the site and the whole idea collapses.
Two sided markets have the unique property that you have two sets of customers who often have opposing desires, and you have to build a solution that keeps both sides happy.
We're contemplating the future direction of the recruitment market and we agree with StackOverflow Careers that the answer is moving over to the model used by dating sites, we've watched carefully what Careers 2.0 have done, and from our perspective they've fallen into precisely the "dating site for guys" trap. They've been so focused in building a site that developers want to use they've missed out doing the customer development on the employer side.
Most companies we've spoken to don't want to hunt through profiles to find the candidates they want, if they wanted to do that they would already be hunting through Linkedin or GitHub (some companies like Google have dedicate recruiters who do this; most companies don't). If a company is hiring for developers it means that their existing development teams are stretched, the last thing they want to do is pull developers off existing projects and have them review profiles of candidates who may or may not even be interested in working for them.
We think we've got a better dating model that will actually serve both developers and employers, and we're going to be rolling it out over the next few months constantly being willing to adapt with what the market actually needs.
Closing note
This article shouldn't be taken as an attack on StackOverflow Careers, the developer job market is a mess and it's becoming dominated by recruiters who are universally hated, we'd much rather see Stack Overflow Careers succeed and crush CoderStack than for both of us to lose out. We're passionate about fixing the developer job market; and we'd much rather see it fixed by one of our competitors than to see it stagnate. In part that's why we're so open, we hope StackOverflow will improve what they're doing. Competition is much more fun that way.
Talking about competition, we actually wrote this article for AppSumo's Lean Startup challenge (if you want to vote for us on twitter you can by clicking here or tweeting with #leanvote63). But we hope it gives an insight into a real world example of lean vs non-lean startup methodologies.
