Why Different Tools Give You Different Grades
You scan your website with one tool and get a B+. You scan the same site with another tool five minutes later and get a 67 out of 100. A third tool says you are in the "needs improvement" category. Which one is right?
They all are — and none of them are. Each grading tool measures different things, weights them differently, and maps results to grades using different scales. Understanding how grading systems work is the difference between chasing a meaningless number and making informed decisions about what to improve.
This guide explains how website grading works under the hood, why scores vary across tools, and how to interpret grades in a way that actually helps you improve your site.
The Anatomy of a Website Grade
Every grading system follows the same basic process, but the details at each step produce wildly different outcomes.
Step 1: Choose What to Measure
The first and most consequential decision a grading tool makes is what to check. A tool that only measures speed will grade a slow-but-secure site poorly. A tool that only measures SEO will give a fast-but-SEO-broken site an A.
Common approaches:
| Tool Type | What It Measures | What It Misses | |-----------|-----------------|----------------| | Speed-only (e.g., PageSpeed Insights) | Performance metrics: LCP, FCP, CLS, TBT | Security, SEO, technology | | Security-only (e.g., SecurityHeaders.com) | HTTP security headers | Performance, SEO, technology | | SEO-only (e.g., basic SEO checkers) | Meta tags, headings, basic on-page signals | Performance, security, technology | | Comprehensive (e.g., PageVital) | Performance + Security + SEO + Technology | Deep crawl analysis, backlink profiles, content quality |
A comprehensive grade that covers multiple dimensions gives you a more accurate picture of overall site health than any single-dimension score. But it also means the grade reflects trade-offs — a site with perfect speed but terrible security will not get an A, and that is by design.
Step 2: Weight the Categories
Once a tool decides what to measure, it needs to decide how much each category contributes to the overall score. This is where scores start diverging significantly between tools.
Consider two hypothetical tools that both check speed and security:
- Tool A weights speed at 80% and security at 20%. A fast site with poor security scores well.
- Tool B weights speed at 50% and security at 50%. The same site scores significantly lower.
Neither weighting is objectively "correct" — it depends on what the tool's creators believe matters most. PageVital uses weights that reflect the priorities of the site owners and marketers in its user base:
| Category | Weight | Rationale | |----------|--------|-----------| | Speed | 30% | Direct impact on user experience and search rankings | | Security | 25% | Foundational protection for visitors and data | | SEO | 25% | Discoverability and search engine compatibility | | Technology | 20% | Stack health, outdated software detection, dependency risk |
These weights mean that no single category dominates the grade. You cannot get an A by excelling in just one area — you need competency across all four dimensions.
Step 3: Score Individual Checks
Within each category, individual checks need to be scored and aggregated. This introduces another layer of variability.
Binary scoring (pass/fail) is the simplest: each check either passes or fails, and the category score is the percentage of passing checks.
Severity-weighted scoring assigns different importance to different checks. A critical security check like SSL carries more weight than an informational check like Permissions-Policy. This means two sites with the same number of failing checks can get different scores if the severity of their failures differs.
PageVital uses severity-weighted scoring:
| Severity | Weight Multiplier | Example Checks | |----------|------------------|----------------| | Critical | 3× | SSL Certificate, HTTPS Redirect, LCP | | High | 2× | HSTS, CSP, X-Frame-Options, FCP, CLS | | Medium | 1× | Referrer-Policy, Image Alt Text, Meta Description | | Low | 0.5× | Permissions-Policy, Lang Attribute, Powered-By Header |
This means failing your SSL check (critical, 3× weight) hurts your security score far more than missing a Permissions-Policy header (medium, 1× weight). The severity reflects real-world impact — a missing SSL certificate exposes all visitor traffic, while a missing Permissions-Policy header is a defense-in-depth measure against a less common attack vector.
Step 4: Map Scores to Grades
The final step converts a numeric score into a letter grade or category label. Different tools use different boundaries:
| Grade | PageVital | Hypothetical Tool X | Hypothetical Tool Y | |-------|-----------|-------------------|-------------------| | A | ≥ 90 | ≥ 95 | ≥ 80 | | B | 80 – 89 | 85 – 94 | 60 – 79 | | C | 70 – 79 | 75 – 84 | 40 – 59 | | D | 60 – 69 | 65 – 74 | 20 – 39 | | F | < 60 | < 65 | < 20 |
A site scoring 82 out of 100 gets a B from PageVital, a C from Tool X, and an A from Tool Y — all from the exact same numeric score. This is why comparing letter grades across different tools is misleading. You should only compare grades from the same tool over time.
Why Your Grades Change Between Scans
Even with the same tool, you might see slightly different results when scanning the same URL minutes apart. This is not a bug — it reflects the reality of how websites work.
Server-Side Variability
Your web server does not respond identically to every request. Load, caching state, geographic routing, and server-side processing time all introduce variability. A speed scan that catches your server during a busy moment will report slower metrics than one during a quiet period.
Third-Party Dependencies
If your site loads scripts from external services — analytics, ads, chat widgets, social embeds — their response times are outside your control. A scan during an outage at your analytics provider will report different performance numbers than one when that provider is operating normally.
CDN Cache State
If you use a content delivery network, scan results can vary based on whether the CDN edge node has your content cached. A "cold cache" request that goes all the way to your origin server is slower than a "warm cache" response served from the edge.
Browser and Network Conditions
Tools that use real browser rendering (as opposed to just fetching headers) are subject to rendering engine variability. Layout calculations, JavaScript execution timing, and paint operations are not perfectly deterministic.
The practical takeaway: do not obsess over a score changing by 2–3 points between scans. Focus on trends over time and significant grade changes (dropping from B to D, for example).
How to Use Grades Effectively
Compare Yourself to Yourself
The most valuable comparison is your site today versus your site last month. Are you improving, stable, or degrading? Letter grades make this immediately visible — going from a C to a B is clearly positive progress.
Avoid comparing your grades to a competitor's grades from a different tool. An A on one platform is not the same as an A on another platform.
Use Grades for Communication, Scores for Action
Grades are excellent for communicating with non-technical stakeholders. "Our website went from a C to a B this quarter" is immediately understood by anyone. "Our LCP improved from 3.2 seconds to 2.1 seconds" requires explanation.
When you need to take action, look at the individual check results behind the grade. The grade tells you something is wrong; the per-check breakdown tells you exactly what.
Track Trends, Not Snapshots
A single scan is a data point. Two scans are a comparison. A series of weekly or monthly scans is a trend. Trends are far more valuable than any individual measurement because they show whether your site's health trajectory is positive or negative.
If your score fluctuates between 82 and 86 week over week, your site is stable. If it has been declining by 2–3 points each month, something is gradually degrading and needs investigation.
Understand Category Contribution
When your overall grade is lower than expected, check which category is pulling the score down. A site with A grades in speed, SEO, and technology but a D in security will have an overall grade around B — that single weak category holds back the entire score.
PageVital's category breakdown makes this visible at a glance. The weighted contribution of each category means:
- Security going from D to A improves your overall score by roughly 20 points (25% category weight × ~30 point improvement)
- Technology going from C to A improves your overall by roughly 5 points (20% weight × ~20 point improvement)
Focus your effort where the category weight and the improvement potential are both high.
The PageVital Approach to Grading
PageVital's grading system is built on three principles that address the most common complaints about website scoring tools:
Transparency. The scoring methodology is fully documented. You can trace any grade back through the category weights, severity multipliers, and individual check results to understand exactly why your site received the grade it did. There is no secret algorithm or proprietary "quality factor" hidden behind the number.
Consistency. The same URL scanned at the same time will produce the same results (within the bounds of server-side and third-party variability discussed above). The scoring formula does not change between versions without notice, so an A from three months ago and an A today mean the same thing.
Actionability. Every check that contributes to your grade has a clear pass/fail status, a severity level, and a description of what it checks. You never receive a low grade without being able to identify exactly which checks failed and what their relative importance is.
Beyond Letter Grades
Grades are a communication tool, not a goal in themselves. An A does not mean your website is perfect — it means the 33 items PageVital checks are in good shape. It does not check your content quality, your conversion funnel, your visual design, or your business logic.
Use grades as a health indicator the same way you would use a medical checkup. Normal blood pressure does not mean you are healthy in every dimension. But abnormal blood pressure is a clear signal that something specific needs attention.
Run a free scan with PageVital to see how your site grades across performance, security, SEO, and technology — then use the per-check breakdown to understand exactly what drives your grade and where the highest-impact improvements are waiting.