Editorial Policy & Data Methodology

Most menu and nutrition sites ask you to trust a number with no account of where it came from. That is a poor deal for the reader, particularly when the figure concerns an allergen or a calorie target someone is actually planning around.

So this page documents our method: the sourcing hierarchy we follow, how we model figures that no brand publishes, how often we re-check, what we do when two sources conflict, and where the honest answer is "we do not know precisely". If you find our method wanting, you will at least be able to say why which is the point.

The 7 Brews Menu is a small independent publication, not a corporate content operation. The team is a mix of regular 7 Brew customers who order across different markets, people who read nutrition documentation carefully for a living, and the engineers who built the calculators.

What that means for our expertise, stated honestly:

• Where we are strong: first-hand ordering experience across many stands and markets, careful reading of published nutrition and ingredient documentation, transparent modelling of customisations, and software that computes consistently rather than by hand.
• Where we are not authoritative: we are not the brand, so we cannot confirm an unannounced recipe change; and we are not clinicians, so we do not give medical or dietetic advice. Our disclaimer states this plainly.
• Why we publish corrections without a byline: some of our best sources work at stands. We credit contributors only when they ask, so that helping us never puts anyone in an awkward position with an employer.

Editorial decisions. What to publish, what to correct, what to hold back, rest with the editorial team alone. Nobody outside it reviews a guide before publication.

Pricing is the hardest data on this site, because a substantial share of stands are independently owned and set prices locally. There is no single national price list to copy, so we build a reference range instead.

1. Collect. We gather menu-board photographs and receipts across multiple states and market types, weighting toward larger metros and known price outliers.
2. Timestamp and locate. Every data point is recorded with its city and date. An undated price is worthless within a few months.
3. Discard outliers. Airport-style, campus and other atypical locations are excluded from the national reference so they do not distort it.
4. Derive a reference. We publish the most commonly observed price for each item and size, not an arithmetic mean, so the figure reflects what most readers will actually see.
5. Require confirmation. A change goes live once two independent, dated sources agree, or one source plus an official announcement.
6. Re-verify. Core items are re-checked on the cycle in section 10, and immediately whenever readers report a shift.

We state throughout that the menu board at the stand is authoritative and that tax is extra. Our pricing disclaimer lists every factor that can move what you actually pay.

Standard menu items use published figures wherever they exist. The interesting problem is everything else: no brand publishes a value for every combination of size, milk, syrup count and topping, and the realistic total runs into the tens of thousands. Those figures are modelled, and here is how.

The model

• Base first. We start from the published value for the standard preparation of the item at a given size.
• Ingredient deltas. Each modification is treated as a documented delta. Per syrup pump, per shot, per milk substitution at the correct volume, per topping.
• Volume and displacement. Size scaling accounts for ice displacement and headspace rather than assuming liquid scales linearly with cup size, which would overstate larger drinks.
• Consistent rounding. Calories are rounded to the nearest 5, macros to the nearest gram, and we round the final figure once rather than at each step so errors do not compound.
• Sanity checks. Modelled figures are compared against published values for similar drinks. A result that is implausibly far from its neighbours is investigated before it ships, not after.

What we tell you about the result

• Modelled figures are presented as estimates, in the same language everywhere on the site.
• Caffeine is presented as a typical value with acknowledged variance. Bean, roast, extraction and steep time all move it, genuinely batch to batch.
• Where a value is uncertain enough that a reader might be misled, we say so rather than smoothing it over.

Every limit here is restated in our nutrition disclaimer, and applies equally to the calorie calculator and the drink customiser.

Allergen information is the one place where being approximately right is not good enough, so we handle it differently from everything else: we constrain the claim rather than stretch it.

• Ingredient-level only. We record whether a documented recipe contains a given allergen. We do not, and cannot, certify a drink as safe for an allergic individual.
• Cross-contact stated everywhere. Shared blenders, steam wands, pitchers, scoops and ice make cross-contact a real possibility that no ingredient list captures. That warning appears wherever allergen data does.
• Precise categories. "Non-dairy" is not "allergen-free". Most non-dairy milks are nut, soy or coconut based, and we label them accordingly.
• Conservative on uncertainty. If documentation for a component is unclear or missing, we flag it as unconfirmed rather than assuming absence.
• Re-checked on any supplier signal. A reported reformulation triggers immediate review, ahead of the normal cycle.

Our locations directory is built from publicly available listings, cross-checked against mapping data and reader reports. Each entry records where the detail came from and when it was last confirmed.

• New openings are added once we can confirm them from a public listing or a dated reader report with evidence.
• Closures and relocations are treated as urgent, a wrong address wastes a real journey.
• Hours are the most volatile field on the site, and we say on the page that they should be confirmed before travelling.
• City and state pages are written from the stands we can verify, not padded with speculative entries to look comprehensive.

Our secret menu database collects unofficial combinations from customers and staff. Because these are by definition not published anywhere official, we apply a different test: not "is this on the menu" but "can this actually be ordered, and does it work".

• Buildable from real components. Every entry must be assemblable from ingredients stands actually carry. Anything requiring a product that does not exist is rejected.
• Multiple independent reports before an item is listed, so we are not publishing one person's private order as a community recipe.
• Ordered as components, not by name. We give the exact order script, because most staff will not recognise an internet nickname.
• Nutrition flagged as less certain than for standard items, since these are modelled entirely from components.
• Clearly labelled unofficial. No stand is obliged to make any of them, and availability depends on local stock.

Being straightforward about this matters, because plenty of sites now publish unreviewed generated text and present it as research.

Where AI is used

• Reader-facing suggestions. Our drink-suggestion features use a language model to phrase a recommendation. These are labelled as suggestions and are not a source of nutrition data.
• Drafting assistance and structure. AI may help draft or restructure explanatory prose, which is then edited by a person.
• Consistency checks. We use tooling to flag inconsistent terminology or figures across pages so a human can resolve them.

Where AI is never used

• Generating nutrition, price or allergen figures. Every number traces to a source in section 3 or to our documented model. A language model is not a source.
• Publishing unreviewed text. No page ships without a person checking every factual claim against source data.
• Inventing citations or sources. If we cite something, it exists and we have read it.
• Fabricating secret-menu items or reader quotes. Both would be straightforward to do and both would be dishonest.

Editorial responsibility rests with the human editorial team regardless of which tools helped produce a draft. "The model wrote it" is not an explanation we would accept from ourselves.

We would rather be corrected today than be wrong for another quarter. Corrections are welcome from anyone. Customers, staff, operators, dietitians, anyone who spots a mistake.

1. Report it. Email the7brewsmenu@gmail.com or use the contact form. A photo of the board or receipt, plus the city, is the most useful thing you can send.
2. Acknowledgement. Within 2 business days, from a person.
3. Verification. We check the report against the hierarchy in section 3 and seek a second confirmation where the change is significant.
4. Correction. Confirmed errors are fixed promptly and the page's date is updated.
5. Disclosure where it matters. If an error could have led a reader to a materially wrong decision. An allergen or a significant nutrition figure, we note the correction on the page rather than fixing it silently.
6. Root cause. Where an error came from our model or process rather than a source, we fix the process too, so the same class of error does not recur.

We do not remove accurate content because a subject dislikes it, and we do not quietly rewrite a page to obscure that it once said something different.

Precision you have not earned is a form of dishonesty, so we use consistent language across the site:

A stated figure

Traceable to an official disclosure or supplier documentation for the standard preparation.

"Estimated" or "approximately"

Modelled by us from documented ingredient values using the method in section 5.

A range

Genuine variability we cannot reduce. Most caffeine values, and figures that differ across preparation methods.

"Typically" or "most stands"

True in the large majority of cases we have verified, with known local exceptions.

"Varies by location"

Set locally by independent operators; we can offer a reference, not a quote.

"Unconfirmed"

Reported but not yet verified to our standard. Shown because it is useful, flagged because it is not settled.

Concretely:

• We are not affiliated with 7 Brew or any competitor. See our non-affiliation notice.
• We buy what we write about. No free product, payment or hospitality is accepted in exchange for coverage or a rating.
• No sponsored content. No paid posts, paid reviews or paid guest posts. If that ever changes, it will be labelled "Sponsored" unmistakably.
• No affiliate links today. If we add them, they will be disclosed at the point of use and in the advertising disclosure.
• Ad placement is structural, not editorial. Ad slots occupy fixed positions and are never woven into a recommendation.
• Competitor comparisons follow published criteria, and we apply the same criteria whichever brand comes out ahead.

When we say a drink is worth ordering, or rank one above another, the basis is stated on the page. Typically that means some combination of:

• Objective measures. Calories, sugar, caffeine, protein, price per ounce which are comparable and checkable.
• Fitness for a stated goal. "Best low-sugar option" is a different question from "best tasting", and we do not blur them.
• Observed popularity where we can support it, described as popularity rather than as quality.
• Declared subjective assessment. Taste notes are opinion, labelled as opinion.

We do not use engineered scores that imply measurement where none happened, and we do not present a preference as a finding.

Editorial quality includes being readable. We aim for plain language over jargon, tables that make sense when read aloud by a screen reader, meaningful link text instead of "click here", and information that never depends on colour alone. Our accessibility statement sets out our conformance target, the measures in place and the limitations we know about.

A methodology page that only lists strengths is marketing. Here are the real limits of what we do:

• We cannot confirm unannounced changes. A quiet syrup reformulation or supplier switch can pass unnoticed until a reader flags it.
• We cannot audit individual stands. Handling practices, calibration and cleaning routines are beyond our reach.
• We cannot verify every price everywhere. There are far more stands than we can visit, and pricing is local.
• We cannot measure your cup. Handmade drinks vary; our figures describe a typical preparation, not the one you were handed.
• We cannot guarantee allergen safety. Structurally impossible for any third party, which is why we say so repeatedly.
• We cannot promise a page is current the moment you read it. Hence the visible dates.

Knowing these limits is what makes the rest of the site useful. A guide that claimed none of them would be less trustworthy, not more.

This page is a commitment, and commitments are worth little without a way to enforce them. If we have fallen short of it, a figure that traces to nothing, a ranking that looks bought, an uncorrected error. Tell us and say so directly.

Editorial standards, corrections and methodology questions

the7brewsmenu@gmail.com

Contact form

https://the7brewsmenu.com/contact

First response

Within 2 business days

Related

About us · Disclaimer & advertising disclosure · Accessibility

Our Community Q&A and comment sections contain content submitted by visitors. User-generated content is clearly separated from our editorially researched content and does not undergo the same verification process described above. Community posts reflect the personal opinions and experiences of individual users, not the editorial team. Always cross-reference user-submitted information with our verified menu, nutrition, and location data before making decisions.