Both Surfer SEO and Frase promise to make a page more competitive, and for affiliate content the difference between them shows up in specific decisions, not in feature lists. How does the tool read a results page full of comparison tables? Does it help you answer a buyer’s question or just add words? Does it push you toward text that costs you trust?
The answers depend on your keyword, your draft and your niche. So instead of reporting scores from a test nobody here ran, this article gives you the exact test to run and the criteria to judge it by, along with where each tool tends to help and where it tends to mislead. If you are deciding whether to pay for an on-page optimizer at all, that is the more useful thing to have.
Why affiliate SERPs ask more of an optimizer than a blog post does
A results page for an informational query is usually a set of similar articles. A results page for “best running shoes for flat feet” or “[product] vs [product]” is a mix: single-product reviews, ranked listicles, comparison tables, forum threads, retailer category pages and brand pages. An optimizer that averages those pages produces guidelines that describe no actual page. A review and a roundup have different structures, and a blend of the two serves neither reader.
Buyer-intent readers also want specific things that term lists do not capture: a verdict near the top, who the product is for, who should skip it, what it costs, and what the writer actually found. A tool can tell you that top pages mention “arch support” 12 times. It cannot tell you whether your page says anything true about arch support.
Trust is the other constraint. Affiliate pages convert when readers believe the writer, and padding damages that belief. A paragraph added to hit a term count reads like filler to a reader who is about to spend money. The anatomy of a high-converting affiliate review shows which sections carry the persuasion, and none of them are there for keyword coverage.
So the question for each tool is whether it helps you cover what the SERP rewards while leaving your page’s evidence and verdict intact.
The test: one buyer-intent keyword, two tools, the same draft
Controlling the draft is what makes this a test of the tools and not of your writing on a given day. Here is the setup:
- Pick one keyword with a commercial modifier, such as “best”, “vs”, “review” or “alternative”. Choose one you actually intend to publish for, and one where you have first-hand material or solid research to draw on.
- Search it yourself first. Before opening either tool, list what the first page contains: how many results are reviews, roundups or comparisons, what each opens with, and whether tables, pricing and verdicts appear above the fold. This is your reference for judging what the tools tell you.
- Write the draft from your own material, without looking at either tool’s guidelines. Save it as a fixed version.
- Run the keyword in Surfer SEO. Record the suggested word count, heading count, term list, and the score your untouched draft receives.
- Run the same keyword in Frase. Record the outline it proposes, the questions it surfaces, the terms it highlights, and the score your untouched draft receives.
- Log every suggestion. For each one, write down whether you accepted or rejected it, and why. “Added because I have a real observation about it” and “added to raise the score” are different entries, and the second list is the cost of the tool.
- Publish one version and keep the log. You will use it later to judge whether the scores meant anything.
Two cautions. First, tool interfaces and the data behind them change, so check what each shows today before building conclusions on what anyone describes from memory, including this article. Second, check current pricing, plan limits and trial terms on each tool’s own site. Most optimizers offer some way to try them, and a trial is enough to run this test on one keyword.
How each tool reads the top results
Both tools analyze the pages currently ranking for your keyword. They differ in what they put in front of you first.
Surfer SEO is organized around its content editor and the guidelines it generates from top-ranking pages: a suggested length, a range for how often to use particular terms, heading targets, and a content score that rises as your draft aligns with those guidelines. It leads with a quantified picture of what ranking pages have in common.
Frase is organized around research and the brief. It pulls the top results, shows their headings, surfaces questions people ask, and assembles an outline you can build from. It leads with what the competing pages cover and how they structure it. It also has a topic score, but the research view is the main thing.
For affiliate keywords, compare each tool against the notes you took in step 2:
- Does it reflect the page types? If the top results are mostly roundups with comparison tables and the tool’s headings suggest a long single-product essay, the averaging has distorted the picture. Neither tool can be assumed to flag this, so look for it.
- Do the headings reveal the recurring sections? Pricing, pros and cons, who it’s for, alternatives and verdict sections that recur across top pages are worth noting. A heading list that shows these patterns is the most useful thing either tool can hand a review writer.
- Are the questions buyer questions? “Is it worth the money?” and “How does it compare to X?” belong on your page. “What is a flat foot?” may belong on an informational page that ranks for a neighboring query, and answering it at length can dilute a buying page.
- Are the terms specs and features, or filler? Product attributes that readers compare are useful. Generic words that happen to appear often on long pages are not.
Neither tool understands intent the way you do after reading the results. Treat their output as data about the pages, and make the intent call yourself.
Where Surfer SEO helps and where it nudges you into bloat
Where it helps. The term list is a fast way to find subtopics you skipped. If ranking pages for a product comparison all discuss warranty, return policy or battery life and your draft says nothing about them, that is a real gap, and a reader may be looking for exactly that. Surfer’s quantified guidelines are also useful when refreshing an existing page, because you can see which sections or terms the current top results have that yours lacks. And if you hire writers, a target score gives them a concrete brief that is easy to check.
Where it pushes toward bloat. A score that rises as you add terms creates an incentive to add terms, and a number is easy to chase. The common failure modes on affiliate pages:
- Repetition to hit usage ranges. Using a phrase eight times because the guideline says so makes sentences stilted, and readers notice stilted writing on review pages first.
- Length from averages. A target word count is derived from competing pages, some of which have long introductions or generic explainers. Matching that count can mean writing 600 words a buyer would skip.
- Terms you have no evidence for. The tool might suggest a feature, a competitor’s name or a use case. If you never used the product that way, adding a paragraph about it means writing something you cannot back up, and the paragraph reads like it.
- Generic paragraphs in the wrong place. Terms often get satisfied by adding an explainer section above the recommendation, pushing the verdict further down the page.
A rule to apply. Add a term or section only when you can say something true and specific about it. If the remaining gaps in the term list are things you cannot support, stop editing. A score below the tool’s maximum with a page you stand behind is a better result than a high score built on padding. The goal of the test is to see how much of your editing was driven by real gaps and how much by the number, and your log from step 6 shows that.
Where Frase helps and where its research-first brief falls short
Where it helps. Frase’s strength is the stage before drafting. Seeing the headings of the top results next to each other, along with the questions people ask, lets you plan a page’s structure in a few minutes, and that structure can be a sensible starting point for a comparison or review. For solo publishers who start from a blank page, a brief that arrives with an outline and a question list saves real time. The question list is also a good source for a short FAQ section near the end of a buyer page.
Where it falls short. A brief built from competitors’ headings tends to reproduce the competition. If every top page has the same seven sections, the tool will suggest the same seven, and your page becomes the eighth identical one. Buyer-intent SERPs reward some structural sameness, because readers expect certain sections, but a page that wins usually has something the others do not: a clearer verdict, a more honest weakness section, a better comparison table, or hands-on detail.
Other limits worth testing:
- The brief does not know your evidence. It cannot tell which sections you can fill with first-hand observation and which you would be padding. You make that call, section by section.
- The questions list is unfiltered. Some questions suit a different search intent. Cut those.
- Built-in AI drafting is tempting. If you let a tool generate sections to hit the brief, you are in generation territory, which has its own trade-offs. The Jasper AI versus Copy AI comparison covers drafting tools; this test deliberately keeps the draft fixed so you are judging optimization alone.
- Scoring may feel secondary. If you rely on Frase mostly for the score, compare how its score responds to your edits against Surfer’s on the same draft. If one tool barely moves when you add a genuinely missing section, that tells you how much to trust it.
Frase gives you a better view of the field before you write. What it cannot do is decide what your page should say that is different.
What the scores did and did not predict once the page was live
A content score measures how closely your draft resembles the pages currently ranking. It does not measure whether readers click through to the merchant, and it does not measure whether the page earns links or satisfies the person searching. Resemblance is a reasonable proxy for topical coverage, and a poor proxy for persuasion.
Rankings depend on many things a score does not see: the site’s authority, backlinks, page age, internal links and how the SERP is changing. One page published after one test is a single data point, and it cannot show that a tool improved your result. What you can do is check whether the score behaved sensibly alongside the real outcomes.
After publishing, keep an eye on a few things over the following weeks and months:
- Impressions and average position in Search Console for the target query, and for the neighboring queries the page picks up.
- Click-through rate from the results page, which depends on your title and description far more than on term coverage.
- Outbound clicks to the merchant and conversions, since these show whether the page persuades.
- Which edits you made for the score and whether the page’s behavior changed around them.
If you test more than one keyword, a pattern becomes visible: whether pages you pushed to a higher score outperformed ones you left at a middling score, controlling for how competitive each keyword was. Until you have several pages, treat any tool’s claim about its score and rankings as a hypothesis. If you need a framework for deciding which numbers deserve attention, which affiliate metrics actually predict profit is a useful companion.
Which one fits which workflow, and when neither is worth paying for
Surfer SEO tends to fit when:
- you draft first and optimize second, and want a score to check the draft against
- you refresh existing pages and need to see what current top results have that yours lacks
- you manage writers and need a measurable brief
- you publish at enough volume that a repeatable check saves time
Frase tends to fit when:
- you plan before you write, and want the structure and questions in front of you at the start
- you are a solo publisher who builds each page from a brief
- you want a fast way to see how top results are organized for a keyword you are new to
- you will use the question list to shape FAQ sections
Many publishers will find one of these workflows is theirs without much testing. If you are unsure, the test above takes an afternoon and your log from step 6 will show which tool’s suggestions you actually used.
When neither is worth paying for:
- You publish a handful of pages. Reading the top ten results yourself and noting their sections, questions and common terms gives you most of what a tool shows, and it forces the reading that improves your judgment about intent.
- Your pages win on first-hand testing. If the page’s value is photographs, measurements, long-term use or a comparison nobody else has done, a coverage score has little to add and may nudge you away from the unusual material that sets the page apart.
- The page ranks but does not convert. An optimizer addresses rankings. If readers arrive and leave, the problem is in the verdict, the offer, the disclosure placement or the page layout, and a case study of fixing affiliate content that stopped converting walks through diagnosing that.
- The subscription would crowd out something with a clearer return, such as hosting, tracking or email tools.
If you do buy one, set a rule before you start editing: no change goes in solely to move the score. That single rule keeps either tool useful for an affiliate site, and the test log will show you quickly whether you were able to keep it.
