Reviews and Promo Sites: What They Can and Cannot Do

A book promo site is not a review, and it is not a substitute for one. It is a paid email placement, with no independent reader judgment behind it. That distinction matters, because the two get treated as interchangeable marketing tools far more often than the evidence supports.

Authors get asked to spend money on the basis of numbers nobody outside the vendor has checked. Peer-reviewed work on reviews exists in some depth, including Chevalier and Mayzlin’s study in the Journal of Marketing Research. Peer-reviewed work on promo newsletters does not exist at all.

This post separates what is documented from what is asserted, then gives you a measurement method you can run on your own book. No figures for named promo sites appear anywhere below, because none are independently verifiable.

The gap between those two sources of evidence explains why this question resists a simple answer. Reviews work through mechanisms researchers have measured directly: sentiment, text detail, and the timing of when a reader posts relative to finishing the book. Promotions work through a channel nobody outside the vendor has opened up. The sections below walk through what Amazon actually states about reviews, why promo site figures fail as evidence, and a method you can run on your own catalogue to get an answer no roundup post can give you.

Key Takeaways

• Amazon’s review rules are written down: incentivized reviews, reviews from people with a close personal relationship to the author, and review exchanges are prohibited under Amazon’s Community Guidelines.
• Consequences are real: Amazon states it may remove reviews, remove content, and suspend or terminate accounts for policy violations.
• Promo site performance data is vendor-supplied: treat averages, download counts, and expected results as marketing copy until an independent source verifies them.
• Rank is a poor measure: count units sold and pages read, since rank moves with other people’s sales as well as yours.
• The test that matters is the baseline: a promotion earned its cost if the fortnight after it sits above the week before it.

What Amazon Actually Documents About Reviews

Some claims in this field are checkable against a published policy document, and some are not. Reviews sit in the first category, which makes them the more reliable half of the book promo sites worth it question, even though reviews and promo sites are different mechanisms entirely.

Amazon’s Community Guidelines prohibit reviews written in exchange for compensation of any kind, including free or discounted products offered on condition of a review. The same guidelines prohibit reviews from people with a close personal or financial relationship to the seller, author, or product. That covers family members, close friends, and business partners, regardless of how the relationship is disclosed.

Review swaps, where two authors agree to review each other’s books, fall under this same prohibition. A review posted in return for anything of value counts, even when the thing of value is another review rather than money.

Amazon’s promotional content policy sets out the consequences it reserves for violations: removing the review, removing the listing, withholding funds, and suspending or terminating the account. These are stated outcomes, not speculation, according to Amazon’s own promotional content policy.

What Amazon does not document is just as worth naming: how it decides which reviews to remove, how long removals take, or how review volume interacts with search ranking. Anyone stating those mechanics with confidence is inferring, not reporting. Advance reader copies remain permitted where no review is required in exchange, and where any relevant relationship is disclosed. Read the guideline itself rather than a summary of it, including this one.

Amazon’s review rules are one of the few parts of self-publishing where the platform has written down what it will do.

Why Promo Site Figures Are Not Evidence

Asking whether book promo sites are worth it usually returns numbers published by the promo sites themselves. A company selling a placement is not a neutral source for the effectiveness of that placement, no matter how the figure is framed.

Almost every average download count, ROI multiple, and expected result circulating in author forums and roundup posts traces back to a vendor page, or to a roundup that copied a vendor page. No methodology, sample size, time period, or independent audit is disclosed alongside these numbers. No peer-reviewed study examines discount newsletters as a category at all. The academic literature covers reviews and ratings, and stops there.

Attribution fails for a specific mechanical reason. Amazon reports only daily aggregate sales, not sales broken out by source. A promotion running alongside organic sales, a newsletter send, a series price change, and seasonal drift all land in that same aggregate number, so no author or vendor can pull the promotion’s effect out of the total on its own. That is the mechanism behind nearly every unverifiable download claim in this industry, and it is why unit tracking across a fixed window is the only defensible workaround.

Practitioner accounts add a related detail. Jim Nelson’s account of his own promo spending notes that most promotion services return no click-level data to the author at all, which leaves the author unable to compute cost per click or cost per download. Treat this as one author’s experience, not a survey.

A pattern that shows up often looks like this: an author books three promotions in one month, sees a rank jump after each one, and credits the placements for a sales lift that a price drop or a new release further along in the series actually produced. By the time the confusion gets sorted out, the quarter’s promotion budget is already spent.

There is a reader-facing consequence too. A large download spike into a mismatched audience can produce reviews from readers who did not finish the book, and reviews written after incomplete consumption are associated with lower subsequent sales, according to research published in Information Systems Research by Lee (2021). The honest position is that the effectiveness of most promo sites remains unmeasured.

How To Measure a Promotion Yourself

You can generate better evidence about whether book promo sites are worth it for your own book than any published figure offers, using a spreadsheet and four weeks. The method below runs on units and pages, not on vendor claims.

Start with a baseline. For the seven days before the promotion, record daily units sold by store from your KDP Reports dashboard, daily KENP pages read, and the price in force. Change nothing else during that week.

Count units, not Best Sellers Rank. Data compiled over the years by AuthorEarnings has repeatedly noted that rank reflects other books’ sales as well as yours, and it cannot be converted back into a precise unit figure. You might notice that a rank number feels reassuring the moment it jumps, even when nothing about your bank balance has changed. That reaction is common, and it is exactly why rank makes a poor measurement tool.

Run one promotion at a time, and avoid simultaneous price changes, category changes, ad launches, or newsletter sends, or the result tells you nothing specific about the promotion itself. A plain spreadsheet works as well as any specialized tool, so use whatever you will actually keep updated.

Watch the decay curve. Log daily units for the fourteen days after the promotion and plot them. The shape matters more than the peak. A spike that returns to the old baseline within three days bought exactly that: a spike.

Compare baselines directly. Average daily units for days 8 through 14 after the promotion against the seven-day pre-promotion average. A higher post-promotion baseline is the outcome worth paying for. Divide the promotion fee by the additional units above baseline across the fortnight to get a cost per retained unit, then decide whether that price is acceptable to you. For series, record read-through separately: units of book two divided by units of book one over the same window.

What the Method Cannot Tell You

This is a before-and-after comparison rather than a controlled experiment, so its limits are worth stating clearly.

• No control group: seasonal and background variation is baked into your result and cannot be removed.
• Single run proves little: patterns appear across three or four promotions, not one.
• Review effects lag: reviews generated by a promotion may appear weeks later, outside the measurement window.

A Worked Example, Clearly Hypothetical

Every number in this section is invented, round, and illustrative. It shows the arithmetic, not a result anyone should expect from a real campaign.

Maybe you have run a promotion, watched the spike, and quietly hoped the higher numbers would hold. Most authors have. The fortnight after the peak is where that hope gets tested.

Take a fiction case, using a public domain title as the stand-in: a Golden Age detective novel such as The Murder at the Vicarage. Suppose the pre-promotion baseline runs at 2 units per day, the promotion day produces 200 units, and days 8 to 14 afterward average 5 units per day. The baseline more than doubled. If the promotion cost is a figure the author can divide into the 21 extra units generated across that fortnight, the cost per retained unit follows directly from there.

Now take a non-fiction case: a nineteenth-century commentary such as J.C. Ryle’s Expository Thoughts on Matthew, where readers arrive by subject search rather than by genre browsing. Suppose the same 200-unit spike occurs, but days 8 to 14 return to 2 units per day. The promotion bought a spike and no lift. The rational conclusion is to stop buying that placement for that book.

The instructive point sits in the contrast: identical peaks, opposite verdicts. Only the fortnight after the peak distinguishes them, which is exactly what most promo site marketing never shows you. The illustration assumes away simultaneous ads, a price change, a review influx, and multiple retailers, so treat it as arithmetic, not as a forecast.

The peak is the cheapest part of a promotion to buy and the least informative part to measure.

What Is Documented, What Is Inferred, and How To Test It

Amazon documents its review policy, its content guidelines, and its terms of service. It documents almost nothing about how browse categories are assigned and refreshed, how the seven KDP keyword boxes function in search, how advertising placements are ranked, or how promotional pricing interacts with visibility.

Treat every confident claim about those mechanisms, including claims repeated for a decade across author forums, as untested until you can name where it was measured and by whom. One adjacent study is worth citing accurately, because it gets misreported often. The eBOUND Canada and Ontario Creates keyword study of 2019 tested ONIX keyword fields supplied through distribution, across roughly 600 Canadian ebooks over one quarter. It found no effect on sales, and a significant increase in discovery, measured as views on Amazon, largest for human-created keywords. It did not test KDP’s seven keyword boxes, which is a different mechanism entirely.

To test an undocumented mechanism on your own listing, change one variable, date-stamp it, hold everything else constant, and measure units and pages across the same seven-day-before and fourteen-day-after frame described earlier. Where the review literature is solid, it supports listing quality over volume: review text carries information that star averages alone do not, according to Chevalier and Mayzlin’s 2006 study. Building a description and a review base that reflect that finding matters more than chasing another promo slot; the book review strategy for marketing success and the

Part of the guide: The Launch: The First Ninety Days