How your offer score works
Every job you see on OpteroAI gets a score from 0 to 100. Here is exactly how we calculate it.
Score breakdown
Your offer score is the sum of five components. Each measures a different dimension of fit. The maximum possible score is 100.
Skills Match
up to 30 ptsHow your skills, experience, and seniority align with what the role requires. We compare your hard skills against the job description, check experience-level fit, and penalize seniority mismatches in both directions. A junior applying for a VP role scores low here. So does a principal engineer applying for an intern position.
Culture Fit
up to 20 ptsWork-life balance expectations, company stability, remote policy, and growth culture compared against your stated priorities. If you ranked work-life balance as a 9/10 and the company has a 2.5 Glassdoor WLB score, this drops.
Salary Alignment
up to 20 ptsWhether the role's compensation range overlaps with your expectations. If your minimum exceeds the role's maximum, that is a hard blocker and the score tanks. Partial overlap gets partial points. When salary data is AI-estimated rather than confirmed from the posting, we note that and reduce confidence slightly.
Historical Success
up to 15 ptsHow candidates with similar profiles have performed at this company and in this role category. Powered by real application outcomes across OpteroAI. If backend engineers scoring 75+ have a 32% offer rate at this company, that data directly feeds into your score.
Stability Signals
up to 15 ptsCompany health indicators: growth trajectory, team size trends, funding stage, recent layoffs, and Glassdoor ratings. A well-funded company with strong reviews and active hiring gets full marks. A company with recent layoffs or declining ratings gets less.
Beyond the fit score
Your fit score tells you how well you match. But even a perfect match does not guarantee an offer. Competition matters.
Hire Probability
Your fit score tells you how well you match. But even a perfect match does not guarantee an offer. Competition matters. A score of 85 at a FAANG company with 500 applicants per role might yield a hire probability of 45. The same score at a 50-person startup with 20 applicants might yield 72. Hire probability is usually 10-20 points lower than the fit score.
Prediction Confidence
How much outcome data we have to back the prediction. Low means few past applications in this role-company combination, so the score is more of an estimate. Medium means reasonable data. High means we have strong historical signal and the prediction is well-calibrated.
Applicant Percentile
Where you rank among the likely applicant pool for this specific role. This factors in skills rarity, company brand (top-tier companies attract stronger pools), seniority alignment, and location. Being in the 90th percentile means roughly 90% of likely applicants are a weaker fit than you.
ATS Pass Probability
Whether your resume would survive an automated keyword filter. We check exact keyword matches, synonyms, and keyword density against required vs nice-to-have skills. If the job asks for Kubernetes and Terraform and you have neither, this drops hard.
The feedback loop
Predictions get better over time. Here is how.
Applications tracked
Every application you make through OpteroAI (or that we detect via Gmail sync) is recorded with its predicted score.
Outcomes collected
When you get an offer, rejection, interview invite, or get ghosted, we capture that result and link it back to the original prediction.
Patterns aggregated
A daily job groups all terminal outcomes by role category and score range. Backend engineers scoring 60-79 have a different success rate than frontend engineers scoring 80-100. We compute these rates from real data.
Future predictions calibrated
When scoring new jobs, the AI receives these historical patterns as context. If the data says candidates like you convert at 25% in this score range, the AI will not predict 70% hire probability. Predictions get honest.
Company intelligence
We track how companies actually behave during hiring. This data comes from real application outcomes across the platform. The more applications we track, the more accurate these signals become.
Response Time
Average days from application to first response. Fast response (under 7 days) signals active hiring.
Ghost Rate
Percentage of applicants who never hear back. A ghost rate above 30% penalizes hire probability by 5-15 points.
Interview-to-Offer Ratio
Of candidates who reach the interview stage, what fraction gets an offer. A ratio above 40% boosts your hire probability.
Hiring Velocity
How quickly the company fills roles. Measured by response speed and active listing count. Fast velocity means they are actively looking to hire.
What moves your score
Score goes up
- Your skills exactly match the required skills listed in the job description.
- Your salary expectations fall within the role's compensation range.
- You have a strong LinkedIn profile with endorsements for relevant skills.
- The company has a high interview-to-offer ratio from past data.
- You are applying to a role that matches your seniority level.
- Your priorities (WLB, growth, stability) align with the company's culture signals.
Score goes down
- Salary mismatch: your minimum exceeds the role's maximum.
- Missing 3 or more required hard skills from the job description.
- Seniority gap: applying for a VP role with 2 years of experience, or vice versa.
- The company has a ghost rate above 30%, which drags hire probability down.
- Location mismatch when the role is on-site and you are remote-only.
- Low ATS keyword overlap between your resume and the job description.
How Gmail access works
Gmail integration is optional. When connected, it automatically detects interview invites, offer letters, and rejections so you don't have to update your application status manually.
What we do
- Read email subject lines and sender addresses
- Classify emails as interview, offer, rejection, or irrelevant
- Link detected events to your tracked applications
- Update your application status automatically
What we never do
- Store full email bodies (only metadata and classification result)
- Read personal, financial, or non-job-related emails
- Sell, share, or use your email data for advertising
- Keep data after you disconnect (tokens revoked immediately)
Technical details
Scopes requested
gmail.readonly (read-only access to email metadata)
Data retention
Classification results kept for 90 days. Email metadata not stored beyond processing.
Compliance
Google API Services User Data Policy (Limited Use). Verified by Google OAuth consent screen.
Revoke access
Settings > Gmail > Disconnect. Or revoke from Google Account permissions.
Prediction accuracy
How well do our scores predict actual outcomes? This table shows real conversion rates by score band, updated daily from tracked applications.
| Score band | Applications tracked | Interview rate | Offer rate | Confidence |
|---|---|---|---|---|
| 80-100 | Collecting | - | - | Low |
| 60-79 | Collecting | - | - | Low |
| 40-59 | Collecting | - | - | Low |
| 0-39 | Collecting | - | - | Low |
Interview rate = reached at least one interview round. Offer rate = received a formal offer. Confidence: High (≥100 applications), Medium (20-99), Low (<20). Data from OpteroAI user outcomes via Gmail sync and manual status updates.
Frequently asked questions
The prediction engine
Every outcome makes the next prediction better.
This isn't a static algorithm. It's a learning system that improves with every application tracked across the platform.
You apply to scored jobs
Focus your time on roles where your odds are highest.
Outcomes tracked automatically
Gmail sync and extension capture detect offers, rejections, and interviews without manual entry.
Patterns emerge across users
Cross-platform data reveals which companies respond fast, which ghost, and what score ranges actually convert.
Your next prediction is sharper
Company intelligence, role patterns, and outcome history feed back into every future score.
The more users track outcomes, the better every prediction gets. That's why OpteroAI improves over time instead of staying static.
How accurate are the scores?
Accuracy improves with data. Early on, scores are AI estimates calibrated with general hiring patterns. As we collect real outcomes (offers, rejections, ghosting) across thousands of applications, the system validates predictions against actual results. We track success rates by role category and score range so you can see exactly how predictive the scores are.
Can I improve my score for a specific job?
Yes. Update your profile with relevant skills. Adjust your salary expectations to overlap with the role's range. Add a tailored resume version that matches the job's keywords. The score recalculates based on your current profile, not a frozen snapshot.
Why do two similar jobs have different scores?
Company intelligence differs. One company might have a 10% ghost rate and fast hiring velocity. Another might ghost 40% of applicants and take 3 weeks to respond. Same role requirements, very different odds of actually getting an offer. Historical success rates at each company also vary.
What is the difference between score and hire probability?
Score measures how well you fit the role on paper. Hire probability estimates your actual chance of receiving an offer, factoring in competition from other applicants, company hiring patterns, and historical conversion rates. A score of 90 at a company that rejects 95% of applicants will have a lower hire probability than a score of 80 at a company that converts half its interviews to offers.
Do you use my data to train models?
Aggregate outcome patterns (like 'backend engineers scoring 75+ have a 28% offer rate') are computed from anonymized, grouped data. Your individual application details are never shared with other users or used to train external models.