You apply for a job.
You make it through several interviews.
Everything seems to be going well.
Then they ask for references.
A few days later, you get rejected.
Most people assume they simply weren’t the best candidate.
But there is another possibility:
One of the references you’re giving employers might be quietly killing your applications.
And you would probably never know.
That’s the business idea.
A reference-checking service for job candidates
The customer isn’t the employer.
It’s the person applying for the job.
You give the service the contact details of a former manager, colleague or other professional reference.
The service performs a realistic reference check and gives you a report showing things like:
- Whether they actually recommend you
- How enthusiastic they sound
- What strengths they mention
- What weaknesses they mention
- Whether they’d hire you again
- Whether they hesitate on important questions
- Anything they say that could hurt your application
You’re basically mystery-shopping your own professional reputation.
The product isn’t really a reference check.
The thing you’re selling is:
Find out what employers are being told about you before it costs you another job.
I would start with software developers
Not because developers are the only people who have this problem.
Because they’re a convenient niche.
Software jobs can pay very well, developers are easy to find online, and losing an offer because of a bad reference can have a very obvious financial cost.
You can always expand later.
But don’t start with:
Reference checking for everyone.
Start narrow.
Why developers might be the perfect market right now
There is another reason I’d start with software developers:
The job market has changed.
A few years ago, experienced developers could get used to recruiters constantly appearing in their inbox.
That isn’t the market many developers are experiencing today.
As of mid-2026, US software-development job postings on Indeed are still roughly 27–30% below their pre-pandemic level, even after a recent rebound.
And that rebound isn’t evenly distributed.
In Q1 2026, 69.3% of software-development postings were for senior-level positions. Indeed also found that 71% of the increase in software-development postings between May 2025 and May 2026 came from senior roles.
The market is especially rough for people earlier in their careers. Indeed found that experience requirements had tightened significantly during the tech hiring slowdown, with junior and standard-level roles hit harder than senior and management positions.
Then there are the layoffs.
Layoffs.fyi has tracked more than 125,000 layoffs across 260+ tech companies in 2026 alone as of August. That isn’t the same thing as 125,000 laid-off developers, but it contributes to a very different feeling around employment in tech than during the hiring boom.
So now imagine you’re a developer in this market.
You’ve sent 80 applications.
Most disappeared into a black hole.
You finally get interviews.
You pass the technical interview.
You speak to the hiring manager.
They ask for references.
Then:
Rejected.
What happened?
Maybe somebody else was better.
Maybe compensation didn’t line up.
Maybe the headcount disappeared.
Maybe you didn’t perform as well as you thought.
Or…
What if your old boss is saying something about you?
You don’t know.
And uncertainty is where this product gets interesting.
Fear can be a very powerful buying motivation
I don’t mean manufacturing fear that doesn’t exist.
That’s shitty marketing and usually a shitty business.
But if the fear already exists, removing uncertainty around it can be extremely valuable.
There is a reason people pay for:
- credit monitoring;
- security cameras;
- medical tests;
- background checks;
- fraud alerts;
- insurance.
Part of what they’re buying is an answer to:
“Is there something bad happening that I don’t know about?”
A reference check has a similar psychological mechanism.
The customer isn’t paying €100 because they desperately want a beautifully formatted reference report.
They’re paying €100 to answer:
“Is this person costing me jobs?”
And the worse the employment market becomes, the more valuable that answer potentially becomes.
When developers had five recruiters messaging them every week, losing one opportunity wasn’t necessarily alarming.
When someone has been job hunting for four months and finally makes it to the end of a hiring process, every hidden variable becomes much more important.
That’s why I think the positioning matters enormously.
I wouldn’t market this as:
Professional candidate reference verification services.
I’d test things like:
Getting interviews but not offers? Check what your references are actually saying about you.
Or:
Would your old boss hire you again? Find out before your next employer does.
Or even:
You check your CV. You practice LeetCode. You prepare for the interview. But have you checked your references?
Now the product connects directly to something developers are already spending enormous amounts of time and emotional energy trying to achieve.
Getting the next job.
This also makes tech YouTubers an unusually interesting distribution channel
There is an entire creator economy built around developers being anxious about getting hired.
YouTube channels make videos about:
- getting your first developer job;
- surviving layoffs;
- FAANG interviews;
- LeetCode;
- CV mistakes;
- why you’re not getting interviews;
- why you’re not getting offers;
- salary negotiation;
- remote jobs;
- career switching;
- surviving the current tech market.
Your product fits directly into that content.
Imagine a creator publishing:
7 Reasons You’re Not Getting Software Engineering Offers
One of those reasons becomes:
“And here’s something most developers never check: your references. One bad reference can potentially affect an offer and you may never know what was said.”
Then:
“There’s a service that will check your references for you before you give them to an employer.”
That doesn’t feel like a random Squarespace ad shoved into the middle of the video.
It’s part of the content.
That’s valuable distribution.
I’d make the affiliate offer aggressive
Suppose the reference check costs €99.
Give the creator €25–€35.
Maybe even more initially.
If fulfillment costs are low enough, I’d much rather give 30% of the revenue to someone who reliably sends customers than spend that money guessing with Facebook or Google ads.
Then approach hundreds of small and medium-sized tech creators.
Not necessarily the giant programming channels.
I’d be especially interested in creators with:
- 10k–200k subscribers;
- job-search content;
- coding bootcamp audiences;
- junior/mid-level developers;
- laid-off tech workers;
- FAANG interview content;
- career-change content.
A creator with 25,000 subscribers whose entire audience is currently trying to get a developer job could be an incredible affiliate.
You could even give each creator a landing page:
Check your references with TECHWITHBOB — 10% off
They get paid.
The viewer gets a discount.
You acquire a customer without paying upfront.
And fear makes affiliate content easier
This is another reason I like the distribution.
The creator doesn’t have to convince someone they need a completely new category of software.
They just have to plant one question:
“Do you actually know what your references say about you?”
If the viewer immediately thinks:
“Fuck. Actually, no.”
the product almost sells itself.
That’s a much stronger trigger than:
Here’s another AI tool that might make you 7% more productive.
And because the developer employment market is currently much less forgiving than it was during the hiring boom, I suspect there are a lot more people for whom that question hits a nerve.
That’s exactly what I’d market-test.
Don’t build software yet
There is almost nothing to build to test this idea.
Create a landing page.
Something like:
Not sure what your references are saying about you? We’ll check.
Offer one manually conducted reference check for perhaps €75–€150.
Then try to sell it.
That’s the test.
Not whether people click a button saying they’re interested.
Not whether Reddit says it’s a cool idea.
Not whether your friends think it makes sense.
Whether someone is worried enough about a reference to give you money to investigate it.
The strongest customer response would be something like:
“I’ve suspected my old manager has been giving me bad references for years. Can you check him?”
Now you have something.
What I’d want to learn first
Before doing anything complicated, I’d want answers to a few questions.
Do people actually worry about this?
Does the fear become significantly stronger after they’ve been rejected at the reference stage?
Do people have specific references they already distrust?
Would they pay €50?
€100?
€200?
Do they want to check one reference, or all of them?
And most importantly:
Does discovering a weak reference cause them to change what they do?
If someone stops using a bad reference and subsequently feels much more confident applying for jobs, you’ve created real value.
Distribution could be very good
This is one of the things I like most about the idea.
The problem naturally produces content people search for:
- Can a former employer give me a bad reference?
- How do I know if I have a bad reference?
- Why do I get rejected after reference checks?
- Can I find out what my old boss says about me?
- What can employers legally say in a reference?
- Should I use my previous manager as a reference?
You could build SEO content around all of these.
Reddit would probably be another obvious channel, especially career and developer communities.
Career coaches and CV/resume services could also become partners.
But one distribution channel I’d test particularly hard is tech YouTubers.
There are huge numbers of developers watching YouTube content about:
- getting developer jobs;
- technical interviews;
- FAANG applications;
- CVs;
- salary negotiation;
- layoffs;
- switching jobs;
- remote work.
The product fits naturally into that content.
A YouTuber could say:
You spend weeks preparing for interviews, but have you ever checked what your references actually say about you?
That’s a pretty strong segue.
Give creators an affiliate share for every completed reference check.
For example, if the service costs €100 and the creator receives €25–€30, the economics could still work well if most of the fulfilment process is standardized.
And unlike paying upfront for sponsorships, affiliates let you test distribution while only paying when you actually make money.
I would especially look for smaller tech career YouTubers with highly targeted audiences rather than immediately chasing giant channels.
Someone with 30,000 subscribers who mainly posts about getting software jobs may be worth far more than someone with 2 million generic tech subscribers.
The first version could be almost entirely manual
The customer fills in a form.
You contact the reference.
You ask a standardized set of questions.
You record the answers.
You send the customer a report.
That’s it.
The first version could literally be:
- Stripe
- Typeform
- Phone
- Google Docs
If people don’t buy that, building a dashboard won’t fix the business.
If it works, automate what repeats
Once you’ve done enough checks, you’ll start noticing what actually needs software.
Maybe that becomes:
- Standardized reference questionnaires
- Automated scheduling
- Call transcription
- AI-generated summaries
- Sentiment and hesitation analysis
- Multiple-reference packages
- Comparison between references
- Suggested references to use
- Career-coach dashboards
- Affiliate management
- White-label versions for job-search services
Or perhaps you discover that customers happily pay €150 for a human-run service and there is no reason to turn it into SaaS at all.
That’s fine too.
The business model should follow the evidence.
The obvious objection
There are legal and ethical questions around how you contact people and how you represent yourself.
That needs to be researched properly for whichever countries you operate in.
I wouldn’t pretend to be a specific company hiring the candidate.
The test needs to be structured in a way that is both credible and lawful.
That’s an implementation constraint, not necessarily a reason the underlying problem doesn’t exist.
And that itself is something I’d investigate before going very far.
Why I like this idea
The best business ideas aren’t necessarily things nobody has thought of before.
They’re often situations where:
- someone has an expensive problem;
- they currently have very little visibility into it;
- solving it produces an obvious result;
- you can test the demand without building much.
This has all four.
Maybe nobody cares.
Maybe developers aren’t the right niche.
Maybe the legal constraints make it impractical.
Maybe people happily pay €200 and tech YouTubers turn out to be an incredible affiliate channel.
I don’t know.
That’s why I’d test it.
Don’t spend three months wondering if it’s a good idea.
Put up the offer.
Find some developers.
Contact some creators.
Ask someone to pay.
Then let reality decide.