SodaPeak Reviews: Seller Figures and Our Review Policy
The SodaPeak seller shows a score of 4.4 out of 5 from 1,359 written reviews.
The seller also lists 6,908 orders. We show these as the seller's own figures, not as verified customer reviews.
Common questions are on the FAQ page.
This page publishes the seller’s review figures as the seller’s figures and publishes none of the eight supplied entries as a verified customer review. Seven of the eight share their wording with at least one other entry, in two clusters, which is the signature of layout copy rather than of eight people typing.
No page on this site emits Review or AggregateRating structured data. We were given no order-linked review record, so there is nothing here for a search engine to show as a star rating, and a star rating we could not stand behind is the one thing on a page like this that travels furthest.
What was supplied, exactly
Eight figures and eight sample entries arrived with the brief. Everything in the table below is the seller’s, reproduced as theirs, with the thing it does not establish beside it.
These are the review figures this website was given. We did not collect them, we cannot audit the system that produced them, and we have no way to link any one of them to an order. Reproducing them as the seller’s own statement is the most we can honestly do with them, so that is what the table does.
| What was supplied | Figure | What it does not establish |
|---|---|---|
| Average score | 4.4 / 5 | Supplied. No method, no date range. |
| Written reviews | 1,359 | Supplied. No link to any of them, and eight samples in total. |
| Orders | 6,908 | Supplied. Not the same thing as buyers, and not verifiable from here. |
| Would recommend | 90% | Supplied. The question that produced it was not. |
| Most common wait before a change | 4 to 8 weeks | Supplied. The only figure here we can hold against a trial length. |
| Most bought pack | 3 bottles | Supplied. A popularity claim, not a price or a result. |
| Most mentioned liked | the daily capsule, and the money-back promise | Supplied as a summary, not as text. |
| Most mentioned disliked | slow start, wanted more in week one | Supplied as a summary, not as text. |
Scroll this table sideways →
The star distribution below is also the seller’s. It sums to 100%, it is consistent with the average score, and neither of those facts makes it independent. A distribution is an output of whatever collected it, and we were not told what collected it.
One figure in that table is worth more than the rest, because it can be held against something real. The most common wait before anyone reported a change is given as four to eight weeks. The best-evidenced ingredient on this label was trialled for twelve weeks, which is longer — so the supplied figure is at least not a claim of an overnight effect.
The eight sample entries, audited
Every entry supplied to this site, with the wording it shares marked. Read the sentences, not the names.
Eight sample entries came with the figures. Below is all eight, unedited, with the repeated wording marked. We have added nothing to them, removed none of them, and written no ninth.
Count the rows rather than trusting any summary, including ours. Three entries carry one sentence pair. Four carry another. One is unique. So 7 of the 8 share their wording with at least one other entry, and the eight entries contain only 3 distinct pieces of text between them.
Note also what the entries do and do not contain. Each carries a first name, an initial, an age, a city, a star count, a pack size and a day number. None carries a date, an order reference, or anything that could be checked by anyone outside the seller. The specificity is in the fields that cost nothing to invent, and absent from the one field that would cost something.
What repeated wording means, and what it does not
Two different sentences appearing under seven different names is not evidence that nobody bought the product. It is evidence about the eight lines we were handed, and the most ordinary explanation is the dullest one: they are layout copy. Somebody needed eight short, similar-length strings to fill a review carousel in a design, and wrote two of them.
That explanation is not an accusation and we are not making one. Placeholder text reaching a live page is a common production failure, and it is the kind of thing a reader can see for themselves in ten seconds once the rows are laid out together, which is why they are laid out together above.
What it does rule out
It rules out treating those eight lines as eight independent reports. Eight people writing about the same capsule do not produce two sentences between them. They produce eight, of different lengths, with different complaints, and at least one spelling mistake.
What it does not rule out
It says nothing about the 1,359 written reviews the same brief claims exist. We were not shown them. They may be entirely real, entirely order-linked and entirely unedited; we have no basis to say otherwise, and no basis to say so either. What we can say is that the sample drawn from them, by whoever drew it, contains only three distinct pieces of text across its eight entries.
The one inference we will not draw
A repeated sentence in a sample is not proof that a review program is fabricated, and we are not going to imply it is. The honest statement is narrow and it is this: what we were given does not support publishing any of it as a verified customer review, so we have not published any of it as one.
Why this site emits no Review or AggregateRating markup
Search engines read a block of structured data on a page and can show what it contains directly in a result — the gold stars, the score out of five, the review count. Two schema types produce that: Review for a single review, and AggregateRating for an average across many. Neither is emitted on any page of this website.
The reason is the same one that keeps the eight entries out of the testimonial slot. Marking something up as an AggregateRating is a statement that the publisher stands behind the figure and the reviews behind it. We do not hold those reviews. We hold a number that was passed to us and eight lines containing two sentences.
What that costs this site, stated plainly
Stars in a search result raise the number of people who click. Declining to emit them is a real cost to a page that exists to sell, and pretending otherwise would be its own small dishonesty. We are stating the cost so that the decision is legible rather than presented as effortless virtue.
What we do emit
Organization and website markup identifying the merchant, breadcrumb markup, article markup on the guides, and FAQ markup on the questions page where every answer is written here and can be read on the page it is marked up from. Nothing that claims a measurement we did not take.
If you see a star rating for SodaPeak in a search result, it came from somewhere other than this domain. That is worth knowing before you weigh it, because the marketplace listing, the ad network and the retailer page are all separate publishers with separate incentives.
What a verified review program actually looks like
This is not a hard standard to describe, and describing it is more useful than complaining about the alternative. Five properties make a review set worth reading, and a set with all five is checkable by the reader rather than only by the seller.
Order-linked
Each review is attached to a real order record, so the person writing it bought the thing. The link does not have to be public; it has to exist, and the publisher has to say that it exists.
One per purchase
A single order produces at most one review. Without this rule the count measures enthusiasm for reviewing rather than number of buyers, and the loudest customer can move an average on their own.
Published unedited
The text goes up as written, including the spelling. Editing for length is where selection starts, and a publisher who edits for length has no principled place to stop.
Dated, and sortable by date
A date tells you whether the set describes the current formula, the current supplier and the current manufacturing run. A review of a reformulated product is a review of a different product.
Negative reviews left in place
The one-star and two-star entries stay, at the same size, in the same list. A distribution with a visible tail is the single strongest signal that nothing is being filtered.
Three further properties are nice rather than necessary: a verified purchase badge that says what it verified, a stated moderation policy naming what gets removed, and a response from the seller shown beside the review rather than in place of it.
None of the five can be established from what was supplied to this site. That is not the same as saying the seller’s program lacks them. It means we cannot tell you it has them, and a page that cannot tell you should say so.
Why a review count is a weak guide in this category
Some of what follows is general to supplements and some is specific to this label. None of it is a finding about SodaPeak, and none of it needs to be, because the argument is about what a review can measure at all.
What is being reported is mostly not measurable by the reporter
The label’s printed claim is “supports healthy muscle”. The blend is called Test Boost. Neither of those is something a person can observe in themselves without a blood test, and the sample entries supplied here report energy, habit and timing — all real experiences, and none of them the thing the name implies.
The effects in question need a control group to see
The best-known trial of saw palmetto, the third ingredient on this panel, was randomised and placebo-controlled, and it found no benefit over placebo for the symptoms it measured. If a controlled trial is required to separate an ingredient from a dummy capsule, an unblinded person reporting on themselves after four weeks cannot do it. That is not a criticism of the person.
Everyone is reporting on a different dose
The side panel says one or two capsules. At one capsule the blend is 600 mg rather than 1,200 mg, so two readers following the label correctly can be taking amounts that differ by a factor of two. A rating that averages across both is averaging across two different products.
And nobody is reporting on a known formula
Eight ingredients share one total and the split is not disclosed. Two ingredients — the bovine testicle tissue and the boron chelate — are absent from the brief and from both marketing images, so a reviewer who read the marketing was not reviewing the product they were holding.
Finally, the category has a documented history on both failure modes. The FDA has published warnings about products marketed as testosterone boosters found to contain an undeclared drug ingredient, and the FTC has settled charges against marketers of androgen supplements over their advertising. Neither is a statement about this product. Both are reasons a star average is a weak instrument here.
The questions to ask of any supplement review set
Use these on any product, including this one, and including any site that reviews this one. They take about a minute and they are answerable from the page in front of you.
- Where did the count come from? A number with no link to the reviews it counts is a claim about a database you cannot see.
- Can you sort by lowest first, and does anything appear? A set with no one-star entries has either a remarkable product or a filter.
- Do the negative reviews describe the product, or the delivery? A tail made entirely of shipping complaints is a tail that has been swept.
- Are there dates, and do they span more than one manufacturing run? Reviews clustered into a single fortnight describe a campaign rather than a product.
- Do any two entries share a sentence? Copy them into a document and sort them. This is the check that took ten seconds on the eight above.
- Does anyone report the thing the name promises, or something adjacent to it? Energy, habit and sleep are adjacent. The panel claim is not.
- What would have had to be measured for the review to be true? If the answer is a blood test, the review is a report of how somebody felt, which is worth reading as exactly that.
- Does the page emit a star rating into search results? If it does, ask who stands behind it and on what record.
The last one is the reason this page exists in the shape it does. A star rating that leaves the page and appears in a search result travels further than the caveat under it, and it travels without the caveat attached.
What we would publish, and what we would need
This page is not closed. If the seller supplies a review export that meets the five properties above — order-linked, one per purchase, unedited, dated, with the negative entries present — we will publish it in full, including the parts that are bad for the product, and we will say on this page when it arrived and what we checked.
Until then, what is above is what we were given. The order desk can be reached Monday to Friday, 9:00am to 5:00pm Eastern Time, and they hold the order records that this website does not. Corrections to anything on this page go to the website inbox.
How this site is edited →What testing was supplied →What the trial lengths were →
References
The sources behind the numbers on this page.
- Bent, Kane, Shinohara et al. — 2006Saw palmetto for benign prostatic hyperplasia — New England Journal of Medicine · DOI 10.1056/NEJMoa053085 · PMID 16467543 · source
- Franco, Trivisonno, Sgarbossa et al. — 2023Serenoa repens for the treatment of lower urinary tract symptoms due to benign prostatic enlargement — Cochrane Database of Systematic Reviews · DOI 10.1002/14651858.CD001423.pub4 · source
- Chinnappan, Rahmat, Mohamed et al. — 2021Efficacy of Eurycoma longifolia (Physta®) on testosterone and quality of life in men aged 50 to 70 — Food & Nutrition Research · DOI 10.29219/fnr.v65.5647 · source
- U.S. Food and Drug Administration — 2023T XTRA Strength Test Booster contains hidden drug ingredient — FDA.gov · source
- U.S. Federal Trade Commission — 1999Androgen Supplement Marketers Settle FTC Charges — FTC.gov · source
- U.S. Food and Drug AdministrationQuestions and Answers on Dietary Supplements — FDA.gov · source
- U.S. Government Publishing Office21 CFR 101.36 — Nutrition labeling of dietary supplements, including the proprietary blend and descending-order requirements — Code of Federal Regulations · source
See the SodaPeak packs
One, three or six bottles. The price is on the seller’s checkout page, because it is the one figure this site was never given — and a figure we were not given is not a figure we will print.
Shipping and the guarantee term are shown at checkout. This site publishes no number it was not able to verify.