Skip to content
VERQEN — home

How we work

Our methodology: no giant report, no random recommendations.

A client should never receive a list of 300 problems with no idea what to do. This page is the whole method, including the rules we use to stop ourselves guessing.
Get a Growth DiagnosticSee a worked investigation
Steps
Five, in fixed order
Analysis layers
Eight, together
Priority system
P0 – P3
Quality gate
7 checks before delivery

The method

Find → Prove → Prioritise → Fix → Measure

Internally this runs as discover, diagnose, prioritise, explain, fix, validate, measure. Each step has an exit condition — we do not move on because time passed.
  1. 01

    Find

    We inspect the data and the website.

    We start with the business, not the keyword list: what you sell, what is worth the most, where leads come from, and what changed recently. Then we collect only the data that question needs — crawl, Search Console, GA4, CRM, logs where they exist.

  2. 02

    Prove

    We connect the problem to evidence.

    Every finding has to be tied to something we can show you. Not “Google probably does not like this”, but “42% of your important URLs return this status while these ones are indexed, and it started on this date”. If the data does not support a conclusion, we say so and tell you what we would need.

  3. 03

    Prioritise

    We work out what actually matters.

    Findings get scored on business impact, confidence and implementation effort, then sequenced by dependency. You should never receive a list of 300 problems with no order. Most projects have three things that matter and a long tail that does not.

  4. 04

    Fix

    We create and execute the right solution.

    Each item becomes an exact change: what to modify, where, and how to test it. Either your developers implement against that specification with us reviewing, or we implement inside the scope we agreed. High-impact changes get staged and reviewed before they ship.

  5. 05

    Measure

    We check whether the change worked.

    Every fix has a metric attached before it ships and a check afterwards, including checking for side effects. When something did not work, the report says so — that is how the next decision gets better.

Analysis

Eight layers, analysed together

Not in sequence. The expensive problems sit between two systems, where nobody owns the boundary — a canonical rule that breaks revenue tracking, a consent change that looks like a traffic drop.
  • L1

    Business

    What is the company actually trying to achieve?

  • L2

    Search

    How are users finding the company?

  • L3

    Website

    Can users and search systems understand the site?

  • L4

    Technical

    Can crawlers access and index the important pages?

  • L5

    Content

    Does the site answer the right commercial questions?

  • L6

    Conversion

    Does traffic become leads and customers?

  • L7

    Measurement

    Can we trust the numbers?

  • L8

    AI Search

    How is the company represented in AI search?

Evidence standard

Facts first.

Every important recommendation needs evidence we can show you. If the data does not support a conclusion, we say so and request what is missing — a guess never becomes a recommendation.

We do not write

“Google probably does not like this.”

We write

“42% of important URLs return a 302 to a canonicalised duplicate, while only the duplicates are indexed. It started on 14 March, the day the new routing shipped.”

What counts as evidence

  • Search Console data
  • GA4 data
  • CRM data
  • Crawl data
  • Server data and logs
  • Page source and rendered DOM
  • Actual search results
  • Competitor comparison
  • Your own business data

Notably absent: intuition, “best practice”, and what worked for a different client. Those can generate a hypothesis. They cannot support a recommendation.

Standard output

Every important issue arrives in this format

Problem, evidence, why it matters, root cause, solution, exact actions, owner, priority, measurement. Nine fields, no exceptions — a finding missing any of them is not finished.

F-014 · Problem

Product pages are canonicalised to their parent category, so individual products cannot rank for their own model queries.

P0· Critical
Evidence
Rendered crawl of 12,480 product URLs: 11,902 declare a canonical pointing at the parent category. Search Console shows the category pages receiving impressions for model-specific queries at position 20+, while no product URL is indexed. The rule was introduced in the release dated 14 March, visible in the crawl diff.
Why it matters
Model-specific queries are the highest-intent demand in this catalogue and currently have no page able to answer them. The category page ranks poorly for them because it is not the right answer.
Root cause
A template-level canonical was added during a duplicate-content cleanup to handle colour and size variants. It was applied to all product URLs rather than only to variant URLs.
Solution
Restrict the canonical rule to variant parameters only. Product URLs self-canonicalise. Variants continue to canonicalise to the parent product, not the category.
Action items
  1. 1.Change the canonical rule in the product template to self-referencing for the base product URL.
  2. 2.Keep variant parameter URLs canonicalising to the base product URL.
  3. 3.Remove product URLs from the category sitemap and publish a dedicated product sitemap.
  4. 4.Add internal links from category pages to the top 200 products by revenue.
  5. 5.Request validation in Search Console for the affected coverage category after release.
Owner
Front-end team, with review by us before release
Measurement
Indexed product URLs (target: >90% of in-stock products within 60 days) and non-brand impressions on model-specific queries, measured against the 28 days before release.

Illustrative finding, shown to demonstrate the format. Not a client result.

Prioritisation

P0 to P3, and what each one means

Priority is assigned on business impact and confidence, then sequenced by dependency. Severity theatre — where everything is critical — makes a report useless.
  • P0· Critical

    Can cause major revenue, traffic or tracking loss. Fix now.

    Example: A noindex tag on a revenue template; conversion tracking not firing at all.

  • P1· High

    Important business impact. Belongs in the current cycle.

    Example: Canonical rules removing a valuable category from search; duplicate conversion events.

  • P2· Medium

    Useful improvement. Schedule it deliberately.

    Example: Internal linking towards commercial pages; consolidating a cannibalising cluster.

  • P3· Low

    Nice to have. Only if it is genuinely cheap.

    Example: Tidying markup that is valid but incomplete; cosmetic metadata improvements.

How priority gets assigned: business impact against implementation effort

Illustrative issue set. High impact and low effort is the do-first quadrant, and it does not need a debate.

DO FIRST — high impact, low effortLow impact, high effort — usually not worth itImplementation effort →Business impact →P0P0P0P1P1P1P1P2P2P2P3P3
  • P0Critical
  • P1High
  • P2Medium
  • P3Low
View the data as a table
Illustrative issue priority scoring
IssuePriorityImpact /10Effort /10
Product canonical rule blocking indexationP09.52
Duplicate conversion event inflating reported leadsP091.5
Redirect map missing high-value URLsP0103.5
Cross-domain tracking not configuredP17.52.5
No commercial pages for comparison queriesP176.5
Facet URLs consuming crawl budgetP16.54
Incorrect company facts in AI search answersP163
Internal linking towards revenue pagesP253
Cannibalising blog cluster on core topicP24.55.5
Core Web Vitals on the category templateP247
Incomplete Organization structured dataP321.5
Rebuild the full analytics dashboard suiteP32.58.5

Source: Illustrative — shows how a real diagnostic backlog is scored

Quality control

Seven checks before anything reaches you

A recommendation that fails any of these goes back, not out.
  1. Is it based on evidence we can show the client?
  2. Is the business impact clear?
  3. Is the action exact enough to implement without interpretation?
  4. Is someone responsible for it?
  5. Is the risk known and stated?
  6. Can we measure the result?
  7. Did we check for side effects?

Reporting

What a report has to answer

Never just clicks, impressions, rankings and traffic. Those describe the past without producing a decision.
  • What happened?
  • Why did it happen?
  • What does it mean for the business?
  • What should we do next?
  • What did we fix?
  • What happened after the fix?

Including the months where the honest answer to the last two is “nothing yet”.

Where AI fits

AI does the heavy lifting. People make the decisions.

Our vision is a consulting company where AI handles the analysis and repetitive work while experienced people make the important business calls. That split is written down, not implied.

AI helps with

  • Pattern detection across large datasets
  • Clustering and classification
  • Summarisation
  • Query generation for testing
  • Competitor analysis
  • Content briefs
  • Task creation
  • Code assistance
  • Report drafting
  • QA assistance

A human decides

  • Business recommendations
  • Migration decisions
  • Deletion and pruning
  • Canonical strategy
  • Redirects
  • High-impact technical changes
  • Any claim about revenue
  • Every client-facing conclusion

The no-guessing rule

If data is missing, we write “we do not have enough evidence to confirm this” and request what is needed. This applies to AI-generated analysis exactly as it applies to ours — a confident-sounding output is not evidence.

Getting started

What happens if you get in touch

Five steps, and you can stop at any of them. No sequence, no pressure.
Get a Growth DiagnosticSee pricing approach

  1. You send us your website and what is happening

    One form, six fields. No call required to get started.

  2. We review the situation

    We look at your site and the public data before we reply, so the first conversation is useful.

  3. We confirm whether we can help

    If we are not the right fit, we say so — and where we can, we point you somewhere better.

  4. We explain what we would investigate

    Specifically: where we would look first and why, based on what we already saw.

  5. You decide whether to move forward

    No pressure and no follow-up sequence. You have the information; the decision is yours.

All ten servicesStart from a problemWorked investigationsWho does the work