Methodology – JustGroupLink Research

JustGroupLink Research — How We Evaluate Keywords, Publisher eCPM, Search Trends and Article Opportunities

Effective date: 21 October 2026
Last updated: 21 October 2026
Website: https://justgrouplink.com
Operated by: Zileth Limited, Umuahia, Abia State, Nigeria

1. Purpose of This Methodology

JustGroupLink Research helps bloggers, publishers, website owners and editorial teams investigate commercially relevant keywords, understand selected historical advertising-performance evidence and develop original article-topic ideas. Our goal is to make research traceable, useful and honest about uncertainty—not to promise particular advertising revenue or search-engine rankings.

This methodology explains the meaning of our figures, the kinds of information we may process, how we examine quality and relevance, how recommendations may be created, and the limits users should understand before relying on a result. It applies to relevant website explanations, private research reports and marketplace research products where those services are available and enabled.

Our methods and software may evolve. The presence of a feature in this document does not by itself mean it is active, fully automated or independently audited. Service-specific screens and disclosures should identify the functions actually available when a report is ordered.

2. Our Identity and Independence

JustGroupLink Research is a research software service operated by Zileth Limited, a company registered in Nigeria. We independently develop our research workflows and editorial recommendations. When permitted third-party information or payment services are used, that does not imply endorsement, partnership or certification by Google, DataForSEO, Paystack, or any advertising network.

We do not represent ourselves as Google Ad Manager, Google AdSense, Google Ads or DataForSEO. We do not require users to supply Google account passwords, advertising revenue-share percentages or private API credentials to upload a manually exported publisher report.

3. The Four Distinct Classes of Information

Different evidence answers different questions. Our research may distinguish:

Advertiser-side keyword information: search CPC, approximate search demand, advertiser competition, paid-bid indicators, country and related terms obtained from authorized research providers.
Publisher-side historical information: advertising impressions, revenue, eCPM, reporting dates and article identifiers voluntarily submitted by a publisher, or lawfully available from other permitted sources.
Editorial and contextual information: topic relevance, intent, audience, location, potential reader questions, originality checks and facts to verify.
Derived research output: filtered keyword lists, comparisons, opportunity scores, topic recommendations and outlines produced by our own processing.

These classes are not interchangeable. In particular, advertiser CPC is not publisher eCPM, and advertising impressions are not page views or Google Search Console impressions.

4. Primary Research Sources

Our intended sources may include:

DataForSEO research services, where connected and purchased, for keyword ideas, related terms, approximate search volume, available CPC fields, competition and supported trend or difficulty indicators;
Google Ads keyword-related data, only if an authorized and appropriately configured integration is available; this is separate from Ad Manager and AdSense publisher reporting;
Manually exported Google Ad Manager or AdSense CSV reports submitted by their authorized publisher for private analysis, and for separately consented research use only when applicable rights permit;
Publisher-supplied article inventories (titles, URLs or post identifiers) for identification and duplicate-topic checks;
Publicly accessible and reputable factual references, consulted as needed when preparing contextual article briefs; and
Our internal research records, subject to accuracy checks, licensing terms and access controls.

We do not use a direct Google Ad Manager or AdSense API connection for publisher CSV analysis. A manually uploaded report remains publisher-supplied material; it is not automatically validated or certified by Google merely because its filename or columns resemble a Google export.

5. Understanding Advertiser CPC

Cost per click (CPC) in our keyword-research context refers to an advertiser-side cost or estimate associated with paid search advertising. Depending on the provider and specific field, the displayed value may be a historical average or a provider-derived bid estimate. It is not a quote that every advertiser will pay and it is not an amount a website publisher receives for an ad click.

We aim to show the source, geographic target, data date, currency and specific metric definition wherever supported. A missing CPC value means the available dataset did not supply a usable figure; it must not be converted into a verified zero, nor classified as a confirmed high-CPC keyword.

High CPC is a discovery signal—not proof of high publisher earnings.

6. What Search Volume, Competition and SEO Difficulty Mean

Search volume usually represents a provider’s estimate of average monthly searches for a keyword over an associated geography and period. It is not a guarantee of visits to any particular article. Exact meanings differ by provider and endpoint.

Advertiser competition describes competition in paid advertising; it is not the same as competition for ranking organically. Organic keyword difficulty, where available, is a separate estimated metric. A keyword can have expensive ads yet be difficult, easy, or uncertain to rank for organically.

We do not present search volume, paid competition, organic difficulty and actual website traffic as though they were the same quantity.

7. What Counts as a Search Trend

We use the word trending only when time-comparable evidence supports a change in search interest or another specified metric. Where data allows, a trend assessment may compare multiple periods within the same provider, keyword definition and geographic scope.

A keyword with a high CPC or high search volume is not necessarily rising. A single recent estimate is not enough to establish an upward trend. Seasonality, incomplete reporting periods, data revisions and one-off news events can distort apparent growth.

When trend evidence is missing or too weak, the result should be labelled trend unverified or insufficient trend data, rather than promoted as a proven trend. The time window and date of observation should be disclosed.

8. The Difference Between Ad Impressions, Page Views and Search Impressions

Three frequently confused numbers measure different events:

Ad impressions: count eligible advertisements recorded as served or displayed under the relevant advertising platform’s measurement rules.
Page views: count viewed pages under the website analytics system’s measurement rules; one page view may produce zero, one or multiple ad impressions.
Google Search Console impressions: count appearances of a property or page in Google Search results under Search Console’s reporting rules; these do not describe advertising impressions on a visited article.

These measurements must be stored and labelled separately. We do not infer the number of website visitors or page views from Ad Manager impression totals alone.

9. What Publisher eCPM Measures

Effective cost per mille (eCPM) describes revenue per 1,000 advertising impressions for the particular revenue and impression metrics being compared:

eCPM = (compatible advertising revenue ÷ compatible ad impressions) × 1,000

For example, $80 in applicable advertising revenue across 2,000 compatible ad impressions produces an eCPM of $40.

Not all Ad Manager eCPM columns are equivalent. A report’s Total average eCPM, Ad Exchange average eCPM or Ad server average eCPM can cover different revenue sources and impression denominators. We aim to preserve these distinctions and will not combine incompatible dimensions or sources into one misleading benchmark.

10. What Page RPM Measures

Page revenue per mille (Page RPM) represents estimated or measured revenue per 1,000 page views, where the relevant page-view count exists:

Page RPM = (compatible page-attributed advertising revenue ÷ page views) × 1,000

For example, $50 in eligible revenue across 5,000 page views yields $10 Page RPM.

An uploaded CSV containing only ad impressions and revenue can support an impression-based calculation; it cannot, by itself, prove Page RPM. We should not fabricate page views, sessions, clicks, Facebook registrations or visitor counts from that CSV.

11. How We Calculate Weighted Historical eCPM

For compatible rows belonging to the same article or clearly defined reporting group, we calculate a combined eCPM using total compatible revenue and total compatible impressions, rather than taking an unweighted average of reported eCPM figures.

Combined eCPM = (sum of compatible revenue ÷ sum of compatible impressions) × 1,000

Illustrative example:

Period	Ad impressions	Revenue	Period eCPM
Period A	100	$9	$90.00
Period B	900	$45	$50.00
Combined	1,000	$54	$54.00

The correct combined figure is $54.00, not the simple average of $90 and $50 ($70). We also check for compatible currency, revenue definitions, time periods, article identifiers and overlapping rows before aggregation.

If the necessary inputs are missing or the denominator is zero, the calculation should be marked unavailable instead of presenting a fabricated number.

12. Historical CSV Intake and Validation

Where enabled, publishers can manually upload reports containing some or all of the following: article URL or supported post identifier, dates or date range, ad impressions, revenue, eCPM, reporting currency and optional country or monetization-channel fields.

A normal validation workflow should check file structure and size, supported column names, missing values, date formats, negative or inconsistent totals, currency and metric types, and repeat submissions. File acceptance does not amount to an independent audit of its underlying authenticity.

The analysis should distinguish a report exported by a legitimate owner from a report whose origins or permissions have not been established. Users remain responsible for submitting only records they are authorized to process. CSV files should not contain passwords, API secrets, payment card data, unnecessary personal information or Google revenue-share percentages.

13. Article Matching and Attribution Checks

Where the uploaded report supplies article URLs, titles, WordPress post IDs or custom key-value identifiers, we attempt to associate historical ad outcomes with the correct article. Differences in tracking setup can create ambiguous results—for example, when a report groups several pages, uses an unmapped identifier or attributes ad activity at a broader inventory level.

We should flag unresolved mappings rather than guessing which article produced the revenue. Users may be asked to confirm a URL-to-ID mapping. Where a report lacks usable article-level identifiers, its results should not be treated as article-specific evidence.

14. Avoiding Duplicates and Misleading Aggregation

We assess whether imported reporting rows overlap in date range, URL, geographic dimension or revenue source. Importing the same report twice must not double an article’s observed revenue or impression count.

We also avoid aggregating incompatible countries, currencies, revenue bases, ad products, reporting definitions or substantially different observation periods without explicit labelling or a defensible conversion method.

A currency conversion, when required, should record the rate source and conversion date. A currency symbol by itself does not establish that the records are comparable.

15. Early eCPM Versus Sustained eCPM

Very high eCPM can occur during an article’s first few impressions. Such a figure may be informative, but a small sample is less stable and may not represent what happens at higher traffic volume.

Where daily or otherwise suitably time-resolved data is available, we may compare early periods with later periods, assess changes as compatible impressions accumulate, and separate short-lived spikes from more persistent performance. We should display the associated impression counts and reporting dates and apply minimum-evidence thresholds where appropriate.

A high historical eCPM is not a forecast of future eCPM. Advertiser demand, audience mix, device type, country, season, auction conditions, traffic quality and other factors can change.

16. Topic Categories and Keyword Clusters

We may organize articles and keywords into related subject areas such as insurance, housing, personal finance, jobs, relocation, education and other relevant categories. Clustering considers how closely terms relate to a reader’s question or editorial subject, not just shared words.

The system should distinguish broad subject similarity from a truly relevant article opportunity. An article about Canadian renting does not establish that every Canadian finance keyword will earn the same eCPM.

Keyword groups should retain the supporting primary and secondary terms so users can review why a recommendation was made. Malformed, repetitive, misleading, off-topic and unsupported suggestions may be removed or flagged during curation.

17. Keyword Filtering and Quality Gates

Where available, research may be screened for:

Relevance to the requested subject or website;
Presence and definition of an advertiser CPC figure;
Minimum or maximum search volume chosen by the customer;
Search intent (informational, commercial, transactional or navigational), when supported;
Geography and currency;
Keyword duplication and near-duplication;
Data age and the availability of comparable trend observations; and
Optional paid competition and organic difficulty metrics.

Filter thresholds depend on the service, user selections and data availability. A keyword that lacks a CPC estimate may still be useful for general editorial research, but cannot be presented as a verified high-CPC record.

18. Understanding Our Rankings and Opportunity Scores

JustGroupLink may offer separate views such as Highest Advertiser CPC, Best Overall Article Opportunity, Highest Historical Publisher eCPM, Most Sustainable eCPM, and Best Personalized New Article Opportunities—where the required evidence and features are available.

A ranking is a decision-support heuristic, not a prediction, third-party certification, or guarantee of revenue. A high score does not mean an equivalent percentage probability of success. Ranking factors may include relevance, CPC, search demand, trend evidence, paid competition, usable historical data, sample size and freshness, with different factors appropriate to different reports.

No fixed universal formula or weighting is asserted here. The active software version and relevant report should explain the scoring method and thresholds it actually uses. If historical publisher data is absent, a publisher-performance ranking should not be fabricated.

19. Personalized Recommendations Using a Publisher’s Own History

Where private analysis is enabled, a publisher may request a comparison between their own historically strong categories and current keyword research. For example, the system may investigate new location-specific questions related to a category that previously showed sustained ad performance.

The resulting recommendation should identify the supporting historical category and independent keyword indicators while distinguishing an observed past outcome from an untested new article. A publisher’s own history is not proof that a different page, country or audience will behave identically.

These personalized outputs remain private to the authorized account unless the owner takes a separate, permissible action to share them.

20. Optional Shared Research Contributions

A publisher’s original uploaded report is private by default. Using it for private analysis does not automatically authorize contribution to shared commercial research. A separate, affirmative permission may be offered for specifically described derived uses, with participation eligibility dependent on applicable law, platform-provider terms and the publisher’s own rights.

Where legally permitted and activated, contribution-eligible patterns may inform independent research into fresh keywords and original article opportunities. Other customers should not receive a contributor’s raw CSV, identifiable earnings, website account details or revenue-share percentages. Small-sample and re-identification risks should be assessed before presenting any aggregate claim.

A 15% contribution-related purchase discount may be offered under the separate Data Contribution Policy. No discount or permission is automatic merely because someone uploads a file. Withdrawal requests and future-use limitations are governed by the applicable privacy and contribution terms.

21. Generating Original Article Topics and Briefs

Where supported, our editorial workflow can turn qualified keyword clusters and permitted contextual signals into proposed titles and article briefs. A report may include the intended audience, country and supported subnational location, financial or practical angle, primary/secondary keywords, suggested headings, FAQs, research questions and facts requiring verification.

A city or province should be included only when geographically relevant evidence supports that angle, or clearly identified as an editorial hypothesis for further research. A proposed title’s inclusion of a year does not verify that its laws, prices or services are current.

Generated topics are starting points for responsible editorial work, not fully fact-checked articles or guaranteed high-earning pages. Claims involving insurance, housing rights, immigration, medicine, taxes and financial products require appropriate current primary sources and, where necessary, qualified professional review.

22. Freshness, Versioning and Corrections

Information can become outdated after collection. Where possible, reports should identify the retrieval date, the provider’s underlying observation period (if known), currency, country, and software/report version.

When newer information materially changes a recommendation, the updated record should be marked as newer rather than silently changing the historical facts. Corrections to imported rows should preserve sufficient audit history to explain calculations while respecting applicable retention and deletion rules.

Previously downloaded reports may differ from newly refreshed research. Users should verify that they are relying on a sufficiently current report for their intended purpose.

23. What the Research Cannot Establish

Even a carefully researched recommendation cannot, on its own, establish:

A specific AdSense, AdX or Ad Manager eCPM that a new article will earn;
Guaranteed publisher revenue, profit or return on advertising spend;
A particular Google organic ranking, impression count or traffic volume;
That a high keyword CPC will translate to high publisher payment;
That historical results from one publisher, audience or period apply to another; or
That a generated legal, health, insurance or financial statement is authoritative without verification.

Results may be affected by errors in the underlying source data, changing markets and the limitations of automated classification. We disclose material uncertainty rather than claiming perfect research accuracy.

24. Research Requests, Credit Charges and Failed Results

JustGroupLink’s proposed service model uses research credits, where 100 credits represent $1 of nominal service purchasing value; these are not withdrawable cash. A customer should receive a maximum quote and description of the expected report before approving a paid operation.

Credits may be reserved and then charged when the contracted report is delivered, with unused reserves restored. A failed promised high-CPC report and a valid market-feasibility report showing weak demand are not necessarily the same outcome. The report’s stated deliverable criteria determine the appropriate treatment.

Current charges, contribution discounts, refunds and customer rights are controlled by the live checkout terms and the separate Refund and Credit Policy and Research Credits and Failed-Search Guide. This methodological explanation does not itself establish that every proposed credit feature is active.

25. Data Security and Responsible Access

We aim to protect publisher-submitted information through role-based access, secure handling of uploads, authentication, appropriate logging and restrictions on public access. The exact retention schedules, hosting locations, subprocessors, technical controls and international transfers must be reflected accurately in the live Privacy Policy.

A contributor’s private report must not appear in another customer’s purchased research. Private account access and optional research contribution are separate permissions. Users are responsible for securing their own login credentials and uploading only information they have the right to submit.

26. Methodological Review and Continuous Improvement

Our intended quality-review process includes examining missing data, incompatible metrics, duplicate imports, unusually high eCPM samples, ranking anomalies, weak keyword relevance, unsupported geographic claims and customer-reported errors.

Research methods may change when we improve validation, introduce new supported data sources or adjust scoring methods. We should disclose material changes in methodology and ensure the research claims shown to customers describe the software actually deployed, not merely planned features.

27. External References and Definitions

For independent explanations of key measurements and provider fields, please consult:

These sources are provided for context. Google and DataForSEO are independent third parties and do not endorse our methodology merely because their documentation is cited.

28. Questions, Disputes and Research Corrections

If a keyword, CPC value, historical eCPM calculation, article mapping, currency, geographic classification or report description appears inaccurate, please contact us with the relevant report identifier, the issue and any non-sensitive supporting details.

We will review reasonable correction requests. For privacy and security, never send Google login credentials, API passwords, payment-card authentication codes or unnecessary personal information in a support message.

Company: Zileth Limited
Business location: Umuahia, Abia State, Nigeria
Research website: https://justgrouplink.com
Research and customer support: info@justgrouplink.com
Administrative and privacy enquiries: admin@justgrouplink.com

Effective date: 21 October 2026
Last updated: 21 October 2026

This Methodology should be read with our Privacy Policy, Terms of Service, Disclaimer, and Data Contribution Policy.