A high commission can make an offer worth investigating. It cannot tell you whether the product fits your audience, whether the program will attribute a sale to you, or what happens when a customer wants a refund. A useful score makes those questions visible before you put your reputation or budget behind an offer.
OfferRanker’s v1 score is an editorial assessment out of 100, based on public evidence available when an offer was checked. It is not a conversion rate, an earnings forecast, a safety certification or a substitute for joining a program and reading its current terms. The total helps you make a shortlist; the six-part breakdown helps you decide what needs a closer look.
This guide explains the rubric, how to gather evidence without filling gaps with guesses, and how to build a repeatable comparison. It also includes two hypothetical offers and a worksheet you can copy for your own research.
Start with the decision, not the payout
Before collecting scores, write down three things: the audience you can actually reach, the problem your content helps them solve, and the action you want them to take. A tool for experienced business owners may have an attractive commission but be a poor recommendation for an audience looking for a free beginner resource. A lower-payout product can be the more useful choice if it solves the right problem and its terms are understandable.
Then set your non-negotiables. These are personal decision rules, not extra OfferRanker score components. For example: “I need to confirm that my email channel is permitted,” or “I will not recommend a recurring subscription unless the renewal price is clear to the buyer.” If an essential condition is unknown, pause rather than using a strong total to make the question disappear.
Use the same standard for every candidate. A score should make comparisons easier, not excuse different evidence requirements for the offer with the largest commission.
The six parts of the OfferRanker score
The published methodology assigns maximum points to six dimensions. The part scores are added; the maximum possible total is 100. The weights reflect an editorial framework, not a formula for expected return.
| Dimension | Maximum | Core question |
|---|---|---|
| Commission and public EPC potential | 25 | What is actually payable, and what earnings evidence is public? |
| Price point and funnel depth | 15 | What does the buyer pay, and what else is presented? |
| Refund policy and terms | 15 | Are material customer and affiliate conditions clear? |
| Sales page and funnel quality | 20 | Can a buyer understand the offer and next step? |
| Vendor and creator track record | 15 | What relevant history can be verified? |
| Public market-demand signals | 10 | Is there observable interest relevant to the offer? |
The rubric is editorial. The methodology does not publish a point deduction for every missing fact or a mechanical conversion from evidence to points. A reviewer must explain the judgment, keep each score within its maximum, and apply the same reasoning across comparable offers. The guidance below helps you inspect that judgment; it does not create an official additional scoring formula.
1. Commission and public EPC potential: 25 points
Find the program’s current published commission and identify the eligible transaction. Is the advertised amount a percentage or fixed payment? Does it apply to the initial purchase, renewals, a particular plan, or a qualified action? Are taxes, discounts or other amounts excluded from the commission base? What must happen before a commission is approved and paid?
Record the exact wording and check whether the offer is product-specific. A marketplace’s general referral program may not have the same terms as the individual product you intend to promote. A generous headline is less informative if eligibility, approval, reversal or payment conditions remain unclear.
Earnings per click (EPC) is generally earnings divided by clicks for a defined period and population. PartnerStack’s EPC definition illustrates that calculation, but a reported figure must still be read with its own source’s definitions. Is it based on all clicks or a filtered group? Which period, traffic, offers and earnings status does it cover? A seller’s figure is a seller claim unless independently substantiated; it is not your forecast. If public EPC is not disclosed, write not disclosed. Do not turn a commission percentage into an assumed EPC or conversion rate.
2. Price point and funnel depth: 15 points
Start with the buyer’s actual entry price and billing interval, not just the price used to calculate an affiliate commission. Note any verified setup fee, trial condition, renewal price, relevant upsell or additional purchase presented in the journey. Identify what you could inspect and where the visible journey stopped.
A higher price can mean a larger payment per eligible sale, but it can also represent a larger commitment for the buyer. More steps may provide useful choices or introduce costs and confusion. Neither a high price nor a long funnel automatically deserves a higher score. A public page that says an offer has “extras” without showing their terms is not evidence of a specific upsell price or commission.
Consider fit alongside the facts: can your audience reasonably use the product, and does the purchase make sense for the problem they are trying to solve? That fit is your decision context, not proof that the product will sell.
3. Refund policy and terms: 15 points
Locate the customer refund policy and the affiliate conditions separately. Look for the eligibility window, how a customer requests a refund, material exclusions, and what the affiliate agreement says about refunded or reversed transactions. Record which terms apply to the particular product and purchase. A platform policy and a seller’s policy can differ in scope; if their interaction is unclear, ask the program which written terms govern.
A refund policy describes conditions; a refund rate is an observed measure. One cannot be inferred from the other. If a refund rate is not publicly disclosed, say so. Do not assume a clear refund window means few customers use it, or that a restrictive policy proves many do.
Under the published methodology, undisclosed customer refund terms receive no credit for those terms. Clearly stated no-refund terms can receive credit for transparency, but this component is capped because of buyer risk; the methodology does not state a numerical cap. The component also considers affiliate conditions, so a missing customer policy should be identified specifically rather than silently treated as an invented refund rate.
4. Sales page and funnel quality: 20 points
Read the page like a prospective buyer. Can you tell what the product is, whom it is for, what is included, what it costs, and what happens after the next click? Are meaningful limitations and recurring commitments easy to find? Does the page provide enough context to distinguish a concrete offering from a vague promise?
Check the visible journey on the devices and locations available to you. Note broken links, unclear plan differences or material information hidden behind an inaccessible step. If a checkout, follow-up sequence or member area cannot be inspected, label it not verified. Do not claim it is absent or assume it is strong.
This is an editorial assessment of the visible experience, not a test of conversion. Attractive design, persuasive copy, testimonials on the seller’s page and scarcity language are not independent proof of results. Do not purchase, submit personal data or create a test transaction unless you have a reason and permission to do so.
5. Vendor and creator track record: 15 points
Identify the business or creator behind the exact offer. Look for a consistent public identity, accessible support and independently verifiable information relevant to the product’s history. Separate an official biography or seller-published testimonial from an independent source. A profile can establish what someone says about their experience; it does not automatically verify the claim.
Consider relevance as well as quantity. A creator’s long public history in one field does not establish expertise in every new offer. An independently documented history can support confidence in a particular fact, but it cannot guarantee future performance or customer support.
Thin public evidence calls for a conservative assessment, not an accusation. Record “not independently verified” where appropriate. A new or small creator may simply have less public history to examine.
6. Public market-demand signals: 10 points
Look for observable interest in the problem or category, using sources relevant to the audience, location and time period you care about. Search activity, public discussions and other visible signals can help show that people are asking about a problem. Record the query or source, what was actually observed, and why it is relevant. Separate category interest from demand for this particular product.
For example, Google explains that Trends data is sampled and normalized to relative interest, rather than absolute search counts. A rising graph is not a count of buyers or a forecast of sales. Small or one-off spikes require particular caution, and a low reading does not prove nobody searches.
An audience description alone is not a demand signal. Under the methodology, if no public signal can be established, this component receives zero. Vendor-published numbers remain attributed claims, not independently verified sales. A zero here identifies an evidence gap; it does not prove that no market exists.
A repeatable research workflow
Step 1: Pin down the exact offer
Record the product name, plan or version, seller, program or network, buyer location if relevant, and the URLs for the sales page and affiliate signup or program terms. Check that the buyer-facing product and the affiliate terms refer to the same offer. If they do not, stop and resolve the mismatch before scoring.
Also note the date you checked each page. A “v1 score, last checked” date describes a snapshot. It does not promise that price, program access or terms remain unchanged afterward.
Step 2: Gather the controlling sources
Start with official buyer-facing pages, customer terms, the specific affiliate agreement and any relevant public marketplace listing. Official sources are usually the right place to find what a seller or program currently states about its own offer, although they may conflict or become outdated. Independent sources can help check history or relevant market interest; they are not automatically authoritative about an individual affiliate contract.
For each material statement, capture a URL, date, specific wording or detail, and the page’s scope. A link to a general homepage is less useful than a link to the actual policy or program page. If you can retain a permitted snapshot for your records, it can help you recognize changes later; always revisit the current source before acting.
Step 3: Label evidence and gaps
Use labels that describe what you know:
- Publicly stated: an identifiable source makes the statement. Attribute it to that source, especially if it reports its own results.
- Independently verified: a relevant source separate from the seller supports the specific claim. Record what it verifies, not a broader conclusion.
- Not disclosed: the relevant public material you checked does not disclose the detail. State the scope of the search; do not imply you searched the entire internet.
- Not verified: a claim exists, but the available evidence does not establish it, or a step could not be inspected.
- Conflicting or outdated: sources differ or a dated statement may no longer apply. Preserve both sources and seek current clarification.
“Not disclosed,” “not verified,” and zero are different things. A zero is a score in a defined component; an unknown is a missing fact. Do not convert every unknown into zero automatically, except where the published methodology specifically requires it for a component. Missing information can limit an assessment even when the overall offer is legitimate.
Step 4: Score and explain each part
For every dimension, write the evidence that supports your judgment, the important unknowns, and the part score out of its maximum. Explain why one offer deserves a different score from another on the same dimension. Add the six numbers and check the total. Do not invent fine-grained deductions or imply that the methodology prescribes a precise score for a particular gap when it does not.
If you are reproducing an OfferRanker assessment, use public evidence and the published rules. A private message from a program can inform your own decision, but it is not public evidence simply because you received it. Note who answered, when, what they said and which written agreement applies; seek written clarification if a message conflicts with the terms.
Step 5: Apply your decision gates
Check program approval, permitted locations and traffic methods, attribution and payment conditions, buyer fit, and the accuracy of your intended promotion. A strong score does not establish eligibility. An unresolved deal-breaker can justify hold, ask, or pass, even when the offer ranks well. A low score can reflect scarce public evidence rather than a poor product.
Worked comparison: same total, different decisions
The following offers, figures and observations are entirely hypothetical. They illustrate editorial judgment and arithmetic, not real vendor claims, a prescribed deduction schedule or actual performance. Both have a public signal of relevant interest in this scenario; neither has a disclosed public EPC or refund rate.
| Component | Hypothetical Offer A | Hypothetical Offer B |
|---|---|---|
| Commission and public EPC potential | 14 / 25 | 10 / 25 |
| Price point and funnel depth | 9 / 15 | 7 / 15 |
| Refund policy and terms | 4 / 15 | 12 / 15 |
| Sales page and funnel quality | 13 / 20 | 9 / 20 |
| Vendor and creator track record | 7 / 15 | 10 / 15 |
| Public market-demand signals | 3 / 10 | 2 / 10 |
| Total | 50 / 100 | 50 / 100 |
Offer A: Imagine a course whose public affiliate page states a relatively attractive initial-sale commission and identifies some affiliate conditions. Its buyer-facing material does not disclose customer refund terms, so those terms receive no credit. The visible page explains the curriculum reasonably well, but the public record for the creator is limited. A few relevant public discussions show interest in the problem, not proven sales of this course.
Offer B: Imagine a subscription with a lower stated initial payout, a clearly described renewal price, and accessible customer refund and affiliate conditions. Its sales page leaves several practical product questions unanswered. Its creator has a more verifiable relevant history than A’s. There is a limited public signal of interest in the problem, but no evidence establishing purchases of this subscription.
The scores reflect those invented circumstances. In A’s refund component, the illustrative points relate to the available affiliate conditions, not credit for undisclosed customer refund terms. The total alone cannot decide which is appropriate: a buyer-protection-focused reviewer might investigate B first, while a reviewer whose audience strongly matches A’s curriculum might ask A for the missing policy before proceeding. Either could be unsuitable if the product, price or permitted channel fails a personal gate.
Now change one fact: suppose A’s program sends a clear customer refund policy privately. That may help an individual assess A, but it does not retroactively make the policy publicly disclosed. For an OfferRanker-style public assessment, look for a current public source before treating it as public evidence. If one appears, record the new source, reassess the affected part, explain any score change and update the total and checked date. Do not automatically award a predetermined number of points.
Check the economics without pretending to forecast
Before comparing payouts, make the commission base explicit. Hypothetical example: a product costs $100 and an agreement pays 40% of that full amount for a qualifying initial sale. If all those conditions hold, the stated commission on that sale would be $40. This illustration assumes no excluded amounts; a real agreement may use a different base or conditions. Renewals, refunds and payout timing require separate checks.
For another hypothetical example, suppose a completed period has 600 reported affiliate-link clicks and $120 in settled, attributed commissions after the relevant reversals. Dividing $120 by 600 gives $0.20 settled EPC for that defined period. If the only measured traffic cost was $30, subtracting it leaves $90, or $0.15 per click after that one cost. That is not full profit: it excludes any unmeasured costs, time and taxes. Pending commissions should not be presented as settled earnings.
Neither example predicts future results. Compare equivalent periods, click definitions and commission status, and avoid treating a small sample as a dependable baseline. A network report and a separate analytics tool may disagree; impact.com explains that reporting configuration, contractual referral windows and other factors can contribute to differences on its platform. Check the actual program’s rules rather than assuming a universal attribution model.
Copyable evidence worksheet
Use one worksheet per offer. Fill in the specific fact and its source; leave unknowns visible. Where a field varies by plan or location, record the scope rather than blending conditions.
Offer identity
- Product and plan: ___; seller: ___; program or network: ___
- Buyer audience and relevant location: ___; proposed traffic channel: ___
- Official sales URL: ___; affiliate signup and terms URLs: ___
- Date checked: ___; page or offer version, if identifiable: ___
Economics and terms
- Buyer price, billing interval and verified extras: ___; source and date: ___
- Commission amount, base, eligible action and renewal terms: ___; source and date: ___
- Attribution window and conditions: ___; source and date: ___
- Approval, payment timing, threshold and reversal conditions: ___; source and date: ___
- Customer refund terms and material exclusions: ___; source and date: ___
- Public EPC or refund rate, definition and period, if disclosed: ___; otherwise not disclosed
Evidence and score
For each row, record the evidence, the source/date, what remains unknown and your reason for the part score. The maxima are fixed; the example spaces are not a preassigned score.
| Dimension | Evidence / source / date | Unknowns and rationale | Part score |
|---|---|---|---|
| Commission and public EPC | ___ | ___ | ___ / 25 |
| Price and funnel depth | ___ | ___ | ___ / 15 |
| Refund policy and terms | ___ | ___ | ___ / 15 |
| Sales page and funnel quality | ___ | ___ | ___ / 20 |
| Vendor and creator track record | ___ | ___ | ___ / 15 |
| Public market-demand signals | ___ | ___ | ___ / 10 |
| Total | Add the six parts | ___ / 100 |
Decision record
- Most important audience-fit reason: ___
- Unresolved questions and who can answer them: ___
- Personal decision gate that could change the outcome: ___
- Decision: proceed / ask / hold / pass; reason: ___
- Next review date or change that triggers review: ___
Keep public evidence and private correspondence distinct. If an answer is private, mark it as such and do not publish it as a public fact. If a link disappears, record that it could not be rechecked instead of silently treating the old detail as current.
Rank independently and audit the score
An independent ranking does not mean ignoring commissions: commission terms are one of the six published factors. It means the possibility of earning a commission must not buy an editorial outcome. OfferRanker says vendors cannot pay for a higher score or ranking, although it may earn commissions from some links. Its affiliate disclosure explains the relationship.
To compare your own shortlist independently, research each candidate against the same six dimensions and personal gates. Record the evidence before choosing a winner. Do not award extra points because you have an affiliate link, received a promotional email, or prefer a vendor. Consider which options remain viable even if you receive no commission. If two offers have the same total, inspect their component scores and buyer fit; a tie does not mean their risks are interchangeable. This is a decision aid, not an invented official tie-breaking rule.
A repeatable audit should answer what changed and why:
- Keep the prior record. Save the old part scores, total, source links, scope, checked date and reasoning. Distinguish a score from a personal decision about whether to promote.
- Reopen the material sources. Compare the current plan, price, program, customer policy and affiliate terms with the record. Check that each source still refers to the same offer.
- Mark the change. Identify whether a fact changed, a claim was newly verified, a source disappeared, or the earlier assessment was mistaken. Do not treat a missing page as proof that the underlying policy changed.
- Reassess affected dimensions. A renewal-price change may matter to both buyer economics and the visible sales journey. Explain each affected judgment rather than changing a total without a reason.
- Recalculate and date. Add all six parts, verify that no part exceeds its maximum, and label the new snapshot. Note what remains unknown.
- Revisit the decision. Check personal gates and promotion copy. If a material term is no longer clear, pause the recommendation or seek clarification rather than relying on an old score.
Audit before a campaign, after a material price or policy change, and on a review schedule you can maintain. Those are practical reader recommendations, not a promise that every OfferRanker listing is updated at a particular interval. The more heavily your decision depends on a term, the more important it is to check that term immediately before acting.
Common scoring mistakes
Treating commission as the entire opportunity. A high percentage may apply to a narrow transaction or exclude renewals. Find the commission base and eligibility rules before comparing it with another offer.
Confusing a claim with verification. A seller’s income screenshot, sales count or testimonial may be useful context about what the seller claims. It does not independently establish typical outcomes, sales or your own future earnings.
Treating missing information as proof of misconduct—or proof of safety. An unknown is neither. Make it visible, score conservatively and decide whether it blocks your recommendation.
Assuming that a polished page converts. A coherent page is easier to assess, but design alone does not reveal how often relevant visitors buy. Separate visible quality from measured outcomes.
Mixing programs or periods. Commission terms for another product, an old promotion or a different geography may not apply. Record the exact offer, scope and date with each source.
Letting the total override a deal-breaker. A high total cannot establish channel permission, buyer suitability or current contract terms. Apply your own gates before promoting.
Using category interest as product sales. Questions about a problem may establish a relevant signal of interest. They do not prove that people purchased the particular offer.
Before you publish a recommendation
Read the current program terms, confirm that your proposed channel and claims are allowed, and make sure your description matches what a buyer can actually obtain. If you have not used the product, do not imply firsthand experience. Attribute claims you cannot independently verify and omit outcomes you cannot support.
Make your financial relationship clear where readers will notice it. The FTC’s endorsement guidance says affiliate relationships should be disclosed clearly and conspicuously and discusses placing a disclosure close to the recommendation. For example, when accurate, you might say: “I may earn a commission if you buy through this link.” The right disclosure depends on the format, relationship and applicable rules; check the relevant guidance and program terms. This article is not legal or tax advice, and a disclosure alone does not make an unsupported claim acceptable.
Frequently asked questions
Is a higher score always the better offer?
No. The score summarizes six public-evidence judgments. An offer with a higher total can still be unsuitable for your audience, unavailable in your location or incompatible with your traffic source. Compare component scores and apply your own gates.
What if a program does not publish EPC?
Mark EPC not disclosed and assess the available commission structure and terms without inventing performance. You can ask the program for clarification, but a private answer is not automatically public evidence or a prediction of your results.
Can two reviewers reasonably assign different scores?
Yes. The rubric is editorial and does not assign a published deduction to every fact. Reviewers should still use the same evidence standards, respect the published rules and point maxima, and explain meaningful differences. If a fact is disputed, resolve the source issue before debating the number.
Does a zero demand score mean nobody wants the product?
No. Under the methodology, zero means no public demand signal could be established in the assessment. It is an evidence limitation, not a measurement of all potential customers or product sales.
How often should I check an offer again?
Recheck material terms before promoting and whenever you learn that a price, policy, program or product has changed. Maintain a schedule that fits your use of the offer, and retain the previous record so you can explain any reassessment.
Your next step
Choose two offers for the same audience and complete a worksheet for each. Open the sources, calculate each six-part total, compare the weakest relevant components, and list the questions that would stop you from recommending either offer. If a material answer remains unknown, ask or hold rather than guessing.
For the audience side of the decision, read how to choose offers for your audience. For warning signs that deserve a closer look, see affiliate offer red flags.