Connecting One Million Careers Through Product
Ethiojobs was never a job board. It was a marketplace whose value lived entirely in match quality.
Senior Product & Growth Lead
- Marketplace
- HR-Tech
- Search
- Product-Led Growth
- Talent
- Active candidates
- 1M+Active candidatesRegistered on the platform
- Daily visitors
- 45KDaily visitorsGrown from 5,000
- Organic search visits
- 118.7KOrganic search visitsIn a single month
- Active employers
- 500+Active employersUsing built-in recruitment tools monthly
The central nervous system of the Ethiopian labour market
Ethiojobs sat at the centre of hiring in Ethiopia, connecting employers with a candidate base that grew past one million people.
Being the incumbent created its own problem. Growth had come from traffic, and the legacy architecture was straining under it.
Competitors were entering, and the defensibility of a job board is thin. Listings can be copied. Match quality cannot.
The moat was never the listings. It was how fast the right two people found each other.
A million candidates, and employers still could not find anyone
Roughly 70% of the candidate base were fresh graduates with limited job market readiness, which meant volume without matched supply.
Employers measured the platform on one thing, which was how quickly they could reach a qualified shortlist. Search speed and candidate quality were the entire value proposition.
A large pool of unprepared candidates actively degraded that experience, so scale was working against us.
Adding more candidates would have made the product worse, not better.
We stopped optimising for listings and started optimising for liquidity.
A marketplace is healthy when both sides transact quickly. Every investment was measured against whether it shortened the distance between an open role and the right person.
Match quality over inventory
Precision search was worth more to employers than another thousand profiles.
Fix the supply, do not just route it
If graduates were not job-ready, the marketplace had to make them ready rather than filter them out.
Own the demand entry point
Search dominance for national hiring keywords was cheaper and more durable than paid acquisition.
Five judgement calls that kept an incumbent from being overtaken
Defending a marketplace is not about shipping more features than the challenger. It is about deepening the thing a challenger cannot copy quickly.
- 01
Rebuild search before adding features
Elastic Search on the Kipawa headhunting platform let recruiters filter over 10,000 CVs by skill cluster in milliseconds, turning search from a bottleneck into the product.
- 02
Build a second product to fix the supply side
Rather than accepting that most candidates were not job-ready, we launched Dereja.com as a career-readiness platform with localized training modules.
- 03
Treat SEO as product, not marketing
Owning the "jobs in Ethiopia" keyword set produced 118,700 organic visits in a single month and removed dependence on paid acquisition.
- 04
Give employers tooling, not just a posting form
An applicant tracking system with automated shortlisting made the platform part of the employer workflow rather than a place they visited occasionally.
- 05
Protect uptime as a growth lever
At 30,000 daily visitors, availability was a competitive feature. Architecture work that produced no visible change was prioritised over features that would have.
Dereja.com turned a liability into a pipeline
The graduate mismatch was the largest structural problem in the marketplace. Instead of filtering those candidates out, we built a separate career-readiness ecosystem that prepares them and feeds them back into Ethiojobs as qualified supply. It now supports over 100,000 graduates annually.
Engineering for national scale traffic
The platform sustained 25,000 to 30,000 daily visitors and more than 580,000 monthly page views with near-zero downtime.
Elastic Search enabled millisecond filtering across a candidate base of over one million profiles.
Traffic concentration reached 98.6% local, placing Ethiojobs among the ten most visited websites in the country.
Building the employer side
Recruitment tooling covered job posting, applicant tracking, and automated shortlisting.
More than 500 organisations moved onto these tools on a monthly basis.
The shift mattered because it moved employers from occasional posters to embedded users with switching costs.
Market leadership held and extended
The platform ended the period with a stronger position than it started, against more competitors than it started with.
- Active candidates
- 1M+Active candidatesRegistered on the platform
- Monthly page views
- 580K+Monthly page viewsWith near-zero downtime
- Daily visitors
- 30KDaily visitorsUp from 5,000 at the start of tenure
- Local traffic
- 98.6%Local trafficTop-10 most visited site in Ethiopia
In a marketplace this size, a one percent improvement in search accuracy translates into thousands of hours of national productivity. The numbers on the dashboard were never the point.
Dereja.com was the outcome I value most, because it treated the weakest part of the marketplace as something to fix rather than something to route around.
In a large marketplace, small improvements compound into disproportionate results.
Liquidity is the product. A product leader in a marketplace exists to ensure talent and opportunity meet with as little friction as possible.
Defending an incumbent position is mostly unglamorous work on speed, relevance, and reliability.
Close the loop between hiring outcomes and ranking
I would feed real hiring outcomes back into search ranking so the marketplace learns which matches actually convert, instead of ranking on keyword relevance alone.