CHI Company Investigation · Healthcare · Positive control · Declared consumer champion audit
Mark Cuban Cost Plus Drug CompanyCost Plus Drugs · Public benefit corporation · Dallas, Texas
Declared Consumer Champion Audit · Verdict: supports · Confidence: medium-high
Cost Plus really did engineer the rent out of the price of a generic drug.Then it measured the saving against a price almost no cash customer pays.
Every other company in this index was investigated on the presumption that it might be hostile. This one declares itself the customer’s ally, so CHI inverted the question: does a company that says it is on your side actually build that alignment into its operating structure, or only into its language? For the core generic cash-price model the answer is substantially yes — and it is verifiable to the cent. What survives against it is not a hidden fee or an extracted margin. It is a savings claim, catalogue-wide, anchored to an industry list price rather than to the discounted cash price a customer can actually obtain today.
The central finding
Two things are true at once, and neither cancels the other.
A 60-page live audit of the catalogue separated the price a customer pays from the saving the page advertises. The first survived without a single exception. The second did not survive at all.
The price is honest.
On every one of the 60 generic product pages sampled from the site’s 2,392-URL medications sitemap, the arithmetic closes exactly: acquisition cost + 15% markup + $5.00 pharmacy labour = the displayed drug price, and displayed price + $5.25 standard shipping = the delivered price.
The whole stack — cost line, markup, labour fee, shipping, comparator and the arithmetic between them — is visible before an account exists. No fee appeared anywhere before the sign-up boundary. Labour was $5.00 and standard shipping $5.25 on all 60 pages without exception.
The savings claim is not.
The “Retail Price at Other Pharmacies” figure that powers every “You save $X” module is a genuine average-retail benchmark — against GoodRx’s own average retail price it tracks at a median of 1.10×. It is not invented and it is not, on this evidence, stale.
But measured against the lowest ordinary coupon price a cash customer can obtain the same day, the same figure runs a median 7.69× higher, exceeds it by more than 10× on 24 of 58 pages, and exceeds every pharmacy price GoodRx lists on 46 of 58. The distortion is a property of the benchmark chosen, not of the price charged.
Section 01 · Why this investigation is different
CHI’s first positive control: a company that declares itself on the customer’s side.
Every company page on this site answers the same question in the same direction: where does this business take money, time, attention or optionality from the customer in ways the customer cannot see or refuse? Cost Plus Drugs was commissioned as the inverse test. The brief was explicit that the answer must not be presumed positive, that brand intent is irrelevant if operating behaviour contradicts it, and that the mission cannot excuse a harmful mechanism.
The larger methodological objective was this: can CHI identify structural mechanisms that suppress hostility, in the same concrete way it identifies mechanisms that create it? A lexicon of hostile patterns can only describe what companies do badly. If customer-favourable conduct is nothing but the absence of hostility, CHI has no vocabulary for it, and no way to say that a hostile mechanism was chosen rather than inherited from the industry.
Cost Plus is the right test case because its entire proposition is a mechanism claim, not a service claim. It does not say it is friendlier. It says the price is built differently, and it publishes the construction. That claim is falsifiable on a public page, without an account, at scale — which is exactly what the live verification did.
The result is the first entry in this index where the interesting finding is what a company removed. Cost Plus demonstrates that a substantial share of pharmacy’s customer-hostile machinery is not a fact of the industry. It is a design decision that can be reversed, by a company operating profitably inside the same supply chain, under the same regulator, buying from the same manufacturers. Every incumbent that keeps those mechanisms is now choosing to keep them.
Scope note. This page assesses Cost Plus Drugs as a consumer pharmacy: the cash mail-order channel, the brand storefront hosted inside it, the terms and policies governing both, and the support architecture around them. The Team Cuban Card retail channel, the B2B Marketplace and the separate Dallas 503B manufacturing facility are described only where they bear on the consumer proposition. No numerical CHI, CVI or CFS score is assigned; none was requested and the research was frozen before calibration.
Section 02 · The generic price architecture
A price stack the customer can audit before the company knows who they are.
This is the load-bearing structure of the whole case. On a public generic product page, before an account exists and before any prescriber is involved, the customer sees the acquisition cost of the drug, a fixed 15 per cent markup on that cost, a $5.00 pharmacy labour fee and a $5.25 standard shipping charge — as four separate lines that add up to the number they will be charged.
What the 60-page audit established
What this removes, mechanically
An itemised margin is not a courtesy; it is a structural impossibility proof. A customer who can see the acquisition cost, the markup rate and the fixed fees cannot be subjected to spread pricing, because there is no second price to spread against. Price increases cannot creep invisibly, because any change appears on the page as a changed line. Coupon-dependent two-tier pricing has nothing to attach to, because the FAQ states plainly: “No, we do not accept coupon or discount cards.” There is no membership, no formulary, no prior authorisation, no rebate and no clawback between the displayed number and the charged one.
Two honest limits belong beside that. The acquisition-cost line is self-reported and unaudited — CHI verified that the arithmetic on top of it is exact, not that the base number is what Cost Plus paid. And the module is empty for branded products, which is the boundary examined in Section 08.
Also observed in the data layer and not rendered on the page: every one of the 60 products reported an in-stock status that the page never displays. The dossier catalogues “no stock indicator” as a real friction for customers who pay before availability is confirmed; the live pass establishes that this is a rendering choice, not a data limitation.
Section 03 · The control comparison
Eight of the ten hostile mechanisms in traditional pharmacy are removed or reduced.
The contradiction-search stream ran a control comparison against ten customer-hostile mechanisms catalogued in the incumbent pharmacy model, then refined it against the live first-party captures. The vocabulary is the brief’s: removes means the mechanism cannot operate in the channel at all; reduces means it operates with less force or scope; untouched means Cost Plus neither removes nor adds it; new friction means the removal creates a different, customer-borne cost. Verdicts below are for the cash mail-order channel.
Opaque pricing
Price known only after adjudication. Substantially reduced by the itemised page — but the cost line is unaudited, the comparator is undefined on the page, and brand products have no breakdown at all. Opacity is removed for the margin, not for the anchor.
Spread pricing
A benefit manager billing a plan more than it pays the pharmacy. Cannot operate: in a cash sale there is no third-party payer to spread against.
Rebate rent extraction
Removed for generics, where rebates are negligible and the published formula leaves no room for one. Unknown for brands, which are manufacturer-set programmes rather than cost-plus products.
DIR fees and clawbacks
Retroactive fees levied on the dispensing pharmacy after the fact. No mechanism exists in a cash sale; in the retail-card channel the dispensing fee is published and stated to be retained by the pharmacy.
Copay above cash price, and gag clauses
The cash price is the only price, so there is nothing for a pharmacist to be forbidden from telling you. For members of the plans the FAQ names, an explicit floor applies: the insured price is always less than or equal to the cash price with standard shipping.
Geographic access restriction
Reduced for any mail-capable patient across all 50 states and the territories. The cost of the removal is latency: 2–3 business days to process, then 5–7 in standard transit, with no same-day option and no controlled substances.
Prior-authorisation burden
Removed entirely in the cash channel. The cost of the removal is that no insurer contributes and nothing credits toward a deductible or out-of-pocket maximum — a trade, not a pure gain.
Formulary exclusion and steering
Formulary logic is gone; a catalogue limit replaces it. Cost Plus lists more than 2,300 commonly prescribed generics and a limited number of brands, with no controlled substances. Independent work found roughly 97 per cent of commercially prescribed generic drug-form-strengths available — broad, but a boundary the customer must check first.
Mail-order forcing
Not addressed, because Cost Plus is mail-only in this channel. A patient who needs a medicine today is not served by it. The retail-card channel is the company’s own mitigation and carries a different fee stack.
Discount-card data monetisation
Removed in observed practice: one analytics container, consent-gated, no advertising pixels found, and a banner pledge not to sell or share personal or health information for advertising. Removed in practice, reserved on paper — the privacy policy still keeps commercial de-identified disclosure and interest-based advertising rights.
Read the ledger by family and the pattern is exact. Every mechanism Cost Plus removes is a rent mechanism — a way of taking money or data through a price the customer cannot see or refuse. Every mechanism that survives is a friction mechanism — time, coordination work, uncertainty or legal optionality shifted onto the customer through a process they can see if they look. The three new frictions it introduces are mail latency, catalogue limitation and the loss of insurance credit. That is the finding of the whole investigation compressed into one sentence, and it is why the verdict is supports rather than strongly supports.
Section 04 · Customer-alignment architecture
Ten named inverse mechanisms — CHI’s first vocabulary for engineered non-hostility.
These are not virtues and they are not marketing. Each one is a page template, a stated commitment or a fee schedule that operates on every transaction, independent of anyone’s goodwill on the day — the same test CHI applies to a hostile mechanism, run in reverse. Eight remove rent, set out below as seven findings because two of them operate as one. Two more are real but weaker, because they are practised rather than promised. Every one is stated with the point at which it stops.
Itemized Margin Disclosure
Load-bearing · verified on 60 of 60 pagesThe product page shows acquisition cost, percentage markup, labour fee and shipping as separate lines before checkout. This is the mechanism the rest of the case rests on. It makes two-sided spread opacity impossible in the channel because there is no second side; it makes price creep legible, because any change has to appear on the page; and it converts a claim about fairness into an arithmetic statement a customer can check.
Where it stopsThe acquisition cost is a self-reported, unaudited number. The module is empty for brand products. And its companion — the “Retail Price at Other Pharmacies” comparator sitting directly beside it — is not itemised, not defined on the page and not measured against the market the customer actually shops in. A disclosure mechanism whose anchor is unexplained is a half-mechanism.
Price-Before-Identity · Catalog-First Workflow
Two names, one structureThe full price is visible without an account, and the company’s own instructions tell the customer to check the catalogue before involving a prescriber. Almost every consumer health service inverts this: identity, then eligibility, then price. Here the economics come first and the identity is optional until the moment of purchase, which removes the leverage that comes from having already collected someone’s data.
Where it stopsAvailability is not shown. The customer learns that a drug is listed, not that it is in stock — and the site holds that stock status in its own data layer without displaying it. The workflow then hands the customer the job of carrying identifiers to the prescriber.
Single-Price Rule
One price, regardless of negotiating skill“No, we do not accept coupon or discount cards.” One price for everyone, with no app to download, no card to present and no reward for knowing how the game is played. In a category where the same drug at the same counter can carry four different prices depending on what the customer produces at the till, refusing coupons is a structural choice against price discrimination, not a limitation.
Where it stopsThe retail-card channel is a second single price with its own fee stack, and the company’s own card FAQ concedes that prices between the two programmes will sometimes vary.
Insured-Price Floor
Price protection, stated unconditionallyFor members of the plans the FAQ names, the price “will always be less than or equal to the Cost Plus Drugs non-insurance price with standard shipping”, with copay and cash price shown side by side before the prescription is requested. This is the direct inverse of the copay-above-cash-price mechanism and of the gag clauses that used to enforce it: the customer is not required to know which price to ask for.
Where it stopsAt the edge of the listed plans. Everyone else does their own comparison — and the independent literature is blunt about how often that comparison favours Cost Plus.
Opt-In Recurrence
Auto-refill chosen, not imposedAuto-refill is “entirely optional and requires you to opt-in”, warns by email seven days before processing, can be cancelled up to seven days before the fill, is locked to the card used at enrolment, and does not retry a declined card. Every one of those is the opposite of the default-on, silent-renewal, retry-until-it-clears pattern this index documents across subscription businesses.
Where it stopsAt the contract. The Terms authorise monthly charges with “no additional notice or consent”, so the mechanism is operational rather than contractual — see Section 07.
Cadence-Neutral Fees
No revenue reason to prefer short refillsA fixed per-prescription labour fee and per-order shipping mean a 90-day fill costs the customer less per day than three 30-day fills, and earns the company less. The incentive to churn the customer through short cycles — the engine behind a great deal of pharmacy refill friction — simply is not there.
Where it stopsThe same fixed fees are nearly the entire price of a cheap generic. The mechanism that removes the churn incentive also prices Cost Plus out of the commodity 30-day fill, which is exactly where the live audit found it losing to $4 lists and warehouse-club prices.
Zero-Cost Exit
Leaving costs nothing and requires no conversationNo membership fee, no lock-in, no minimum, self-service account deletion, and no cost to having a prescription routed elsewhere. Exit resistance is one of the most reliably present mechanisms in this index; here there is nothing to resist with, because nothing was ever charged for staying.
Where it stopsExit is free but not costless. A patient who leaves mid-supply still needs a new prescriber contact and bridges the gap alone — the same gap that turns a late shipment into an emergency purchase at full retail.
Tracker Minimalism
Practised, not promisedOne tag container, analytics only, consent-gated, and no advertising pixels found — a materially lighter footprint than the discount-card conduct the FTC penalised in 2023 and than the pixel-litigated pharmacy sites this research compared it against.
Where it stopsThe privacy policy reserves the right to abandon it: de-identified commercial disclosure, interest-based advertising and marketing partnerships all remain on paper, with an email-only opt-out. Conduct is better than the document, which is the reverse of the auto-refill problem and just as incoherent. Cost Plus is also the first pharmacy CHI has assessed, so there is no in-corpus pharmacy baseline to rank it against.
Small-Claim Fee Advancement
Practised, not promisedThe Terms advance arbitration costs for claims of $10,000 or less and contain no mass-arbitration batching clause — better than every incumbent comparator examined in this research.
Where it stopsIt sits inside an otherwise defendant-protective contract: mandatory arbitration with no opt-out, a Texas seat, Dallas-only small claims, a broad liability release and a one-year limitation period. On the evidence it is probably an artefact of how a 2023 document was drafted rather than a designed customer protection, and it is credited at that weight.
Why this matters beyond one company. Each mechanism above is a demonstration that the corresponding hostile mechanism is optional. Before this investigation, CHI could only score the absence of a hostile pattern against a hypothetical alternative. It can now score it against a working one: a company that itemises its margin has proved that itemising margin is commercially survivable, and every company that does not is making a choice.
Section 05 · The two-benchmark savings paradox
The same number is faithful against one benchmark and wildly inflated against another.
Beside the honest price stack, every generic product page carries a second figure — “Retail Price at Other Pharmacies” — and subtracts the Cost Plus price from it to produce a “You save $X” headline. The live audit tested that comparator on 60 pages against same-day controls, using two different definitions of what a customer “would otherwise pay”. The two answers are the finding.
Cost Plus comparator ÷ GoodRx’s own displayed “average retail price” · n = 43 pages
Reading: the comparator is what the FAQ says it is — “an average of retail prices across other pharmacies… sourced from third-party data”. It tracks an independent average-retail figure closely: 36 of 43 within 1.5×, none above 3×. It is not fabricated, and across this sample it is not demonstrably stale either.
Cost Plus comparator ÷ lowest non-membership coupon price at a named pharmacy · n = 58 pages
Reading: against the price a coupon-using cash customer actually faces, the same figure is more than 1.5× on 55 of 58 pages, more than 3× on 48, more than 5× on 39 and more than 10× on 24 — and it exceeds every pharmacy price GoodRx lists on 46 of 58. The exaggeration is inherent to using a list-price average for a savings claim.
And the counterweight, which matters just as much: Cost Plus’s own delivered price beat the lowest obtainable coupon outright on 29 of 58 pages and came within $2 of it on 17 more — 46 of 58 competitive or better. It lost by more than $5 on seven commodity generics, where the fixed $5.00 labour fee and $5.25 shipping are nearly the whole price. So the customer is usually getting a good deal. They are just not getting the deal the page says they are getting.
Bars are drawn as a share of each benchmark’s own sample, so the two distributions are directly comparable.
Two further facts frame the size of the effect. By price tier, the comparator-to-coupon ratio runs at a median of 6.0× across the 30 cheapest pages, 13.0× across the 23 mid-priced ones and 11.3× across the five most expensive — so this is not an artefact of a handful of specialty drugs. And 58 of the 60 sampled pages produced a usable control: three needed normalising from a 60-count default, and two were limited-distribution products for which no pharmacy price table exists at all.
The two benchmark systems, side by side
| Question | Benchmark A — average retail / list | Benchmark B — obtainable cash price |
|---|---|---|
| What it represents | What a pharmacy’s posted, unnegotiated price averages across the market | What a person paying cash can get today with a free, publicly available discount coupon |
| Median ratio to the Cost Plus comparator | 1.10× | 7.69× |
| Sample | 43 pages | 58 pages |
| Worst case in the sample | Under 3× | 219× (imatinib) |
| Is Cost Plus’s figure defensible against it? | Yes — and it matches the company’s own published description of the method | No — the gap is systematic, catalogue-wide and undisclosed on the page |
| Which one does the customer live in? | Almost nobody, for a generic drug in 2026 | Anyone with a smartphone and thirty seconds |
↔ Scroll table horizontally on small screens
What the company says about the comparator, and where it says it
That answer is accurate, and it discloses the staleness risk honestly. Three things about it are still adverse. It is the only place the comparator is defined: the product page never explains it, and the “Learn More About Our Transparent Pricing” modal — which patiently defines Manufacturing, 15% Markup, Pharmacy Labor and Shipping — is silent on the one number driving the savings headline. The third-party source is unnamed. And the disclosure warns about the wrong risk: the audit found drift is not the problem, and that the benchmark type is. Nothing on any surface tells the customer that coupon and discount-card prices are excluded from the average, which is the single fact that would let them read the number correctly.
Stated precisely, because the distinction carries the section: Cost Plus is not inventing a retail price, is not demonstrably publishing a stale one, and is not misstating what it charges. It is choosing a benchmark — the same usual-and-customary construct that discount-card programmes use for their own “save X%” claims — that has not described the cash market for years, and presenting the difference as the customer’s saving.
Section 06 · The audit, in rows
Twelve representative pages from the sixty.
Selected across the full price range and deliberately including the cases that cut against the finding — the commodity generics where Cost Plus loses on price, and the one page where its comparator sits below the obtainable coupon. All figures captured 5 September 2026. Delivered price includes $5.25 standard shipping.
| Drug (strength × quantity) | Cost Plus delivered | Comparator shown | Average retail control | Lowest obtainable coupon | Comparator ÷ coupon |
|---|---|---|---|---|---|
| Atorvastatin 10 mg × 30 | $10.50 | $55.20 | $59.45 | $10.52 | 5.2× |
| Metformin 500 mg × 30 | $10.56 | $20.10 | $22.26 | $4.76 | 4.2× |
| Levothyroxine 50 mcg × 30 | $11.18 | $16.80 | $18.88 | $4.00 | 4.2× |
| Montelukast 10 mg × 30 | $10.96 | $170.10 | $71.02 | $10.76 | 15.8× |
| Tadalafil 5 mg × 30 | $11.22 | $306.90 | $200.79 | $12.06 | 25.4× |
| Solifenacin 5 mg × 30 | $11.32 | $337.50 | $308.07 | $13.50 | 25.0× |
| Aripiprazole 10 mg × 30 | $13.01 | $677.70 | $551.06 | $14.72 | 46.0× |
| Abiraterone 250 mg × 30 | $24.50 | $1,093.20 | $2,451.31 | $31.20 | 35.0× |
| Everolimus 10 mg × 30 | $303.50 | $19,347.00 | $14,246.89 | $805.73 | 24.0× |
| Tofacitinib 5 mg, bottle of 60 | $30.33 | $4,449.00 | $4,818.01 | $3,941.00 | 1.1× |
| Imatinib 400 mg × 30 | $39.75 | $9,657.30 | not shown † | $44.00 | 219.5× |
| Capecitabine 500 mg × 30 | $18.29 | $159.30 | not shown † | $292.20 | 0.5× |
↔ Scroll table horizontally on small screens
† For these drugs the control source displayed only a coupon “standard price” in place of an average retail figure; those rows are excluded from the average-retail benchmark and appear only in the coupon comparison. On the first three rows Cost Plus is the more expensive option: metformin costs $5.80 more than the cheapest obtainable price, levothyroxine $7.18 more — the fixed-fee floor doing exactly what Section 04 said it would. On the last four, Cost Plus is dramatically cheaper than the cash market, and on capecitabine its own comparator is lower than the coupon price, which is the audit finding running in the opposite direction and is reported here as found.
Section 07 · The illustration, not the argument
Imatinib: what a 219× benchmark gap looks like on one page.
The imatinib page is the most visually extreme case in the catalogue, and it is used here only as an illustration. The systemic 60-page audit is the evidence. One striking product page proves nothing on its own; that is precisely why the audit was run.
Why the $9,657 figure is not fabricated
The same control data that produced the $44 coupon price also lists a warehouse-club cash price of $8,755 for the same drug, strength and quantity, and a second club price of $794.99. A list-type price in that region demonstrably exists in the real retail market. Cost Plus did not invent the number, and this page does not say it did.
What the number does is measure the saving against a market almost nobody shops in. A patient who walked into that warehouse club without a coupon really could be quoted something close to $8,755. A patient who spent thirty seconds on a free discount site pays $44. The page offers the first figure as the alternative to Cost Plus and stays silent about the second — and Cost Plus, at $39.75 delivered, genuinely is the cheapest of the three. The company is winning this comparison on the merits and advertising it with a number that overstates the win by roughly two hundred times.
Under CHI’s Fixed Under Pressure principle, a correction here would be credited in full and immediately. Three drafting changes — defining the comparator on the product page, naming the data source, or replacing the benchmark with a current cash-market figure — would close the single adverse rent-family finding in this investigation. None of them requires Cost Plus to spend a dollar more on service or accept a cent less margin.
Section 08 · What did not survive
The friction Cost Plus kept, and the two places its own documents disagree with each other.
A favourable verdict is not a clean sheet, and the contradictory evidence is not softened because the overall finding is positive. Below the product page — after the customer clicks checkout — Cost Plus is an ordinary low-touch mail-order pharmacy, and in two respects it is worse than ordinary: the FAQ points to a telephone that is not there, and the savings module anchors to a price the customer will not be charged anywhere they are likely to shop.
There is no customer-service telephone number
The consumer site does not expose a general customer-service phone line. The FAQ nevertheless instructs, verbatim and still live: “How do I change or cancel my order? Give our customer service team a call right away!” — with no number attached, in the one situation where minutes matter.
Exactly one telephone number appears anywhere on the consumer site: (833) 926-3384, once, in the FAQ under the heading For Providers, described as the line for a prescriber changing a prescription already sent and routed to the company’s pharmacy fulfilment partner. Whether that number is owned by the partner or by Cost Plus and routed onward cannot be established from first-party evidence, and this page does not assert it. Support for patients is a web form with a routing-confirmation step.
The founder has stated the trade openly — the company “won’t be the fastest or highest touch source… a reality that comes with being the low-cost provider” — and CHI credits that disclosure. What is not disclosed is the gap between the trade as described and the trade as delivered: an instruction to call, and nothing to call.
The support chatbot cannot do the urgent thing
A floating help widget now opens an AI assistant whose own opening message reads: “I’m an AI assistant, here to help with quick questions like order status, prescription updates, and FAQs. I’m not a live person, so I can’t place orders or change prescriptions. If you need additional support, please visit our Contact Us page and we’ll connect you with the right team member.”
CHI records this as customer-service friction, not as a new extraction mechanism, and credits the disclaimer as unusually candid. But it does not close the gap above: the one time-critical action the FAQ tells customers to telephone about — stopping an order — is the action the bot explicitly cannot perform, and it routes back to the same web form.
The contract does not make the promise the FAQ makes
The FAQ commits: auto-refill is “entirely optional and requires you to opt-in”; “we will send you an email 7 days before your order is processed”; and “you can cancel your auto-refill order up to 7 days before it gets filled”.
The Terms of Service, still carrying a June 2023 effective date, say something else: the company is authorised to charge the card “and that no additional notice or consent is required”, and for a subscription service such as auto-refill will charge “on a monthly basis until you cancel your subscription”.
Assessed as found: this is a drafting inconsistency in which the contract is weaker than the promise, with no evidence in the record of the contract being applied against the promise. The Terms do incorporate supplemental terms posted on the service, which arguably pulls the FAQ commitments in; the two texts are simply never reconciled, and whether the described flow operates as promised could not be verified without an account.
Fulfilment is a one-way pipeline
Processing takes 2–3 business days before 5–7 business days of standard transit; the company never states the combined door-to-door figure, and expedited shipping does not expedite processing. There are no returns, cancellation is not guaranteed once placed, stock is confirmed after payment rather than before, and the delivery address cannot be changed.
The complaint record is consistent with that design: of 47 complaints closed at the Better Business Bureau over three years, 21 concern delivery, and individual complainants cite elapsed times of 10, 11, 12, 14, 15, 21 and 40 days. Public review venues run negative — 2.5 on 114 Trustpilot reviews, 1.54 on 79 BBB reviews — while the positive reviews that do exist are almost monothematic about price.
The research concluded these are meaningful friction mechanisms that do not negate the economic alignment of the generic pricing model. They are also where the cost lands on a specific group: the complaint corpus skews older, chronically medicated and paying cash — the constituency the model most claims to serve.
On the review evidence generally: these are self-selected, negativity-skewed venues, the BBB profile is structurally a complaint channel, and against a customer base the founder has described in the millions the absolute counts are vanishingly small. They establish that the same failure modes recur; they cannot establish how often they occur per order. Reddit — probably the largest body of Cost Plus commentary, and likely less negative — was unreachable during the research window, which biases the complaint share upward.
Section 09 · Where the model ends
The branded storefront is a different structure, and that is the point.
The strongest alignment finding in this investigation applies to one specific thing: the core generic cash-price model. The branded products sold on the same site do not run on that architecture, and the Eliquis storefront is the clean example. Naming the boundary is not an accusation. It is what stops a favourable finding from being over-claimed.
Inside the model
Generic cash price · cost-plus-15 architecture- Acquisition cost, 15% markup, $5.00 labour and $5.25 shipping shown as separate lines.
- Arithmetic reconciles to the cent on every page audited.
- Price fully visible before an account exists.
- One price for everyone; no coupons accepted, no membership, no formulary.
- Insured-price floor for members of the listed plans.
- “You save” comparator present — and carrying the finding in Section 05.
Outside it — the Eliquis storefront
Manufacturer-sponsored direct-to-consumer cash programme- $345.00 for either the 2.5 mg or the 5 mg bottle of 60, described on the page as a “direct to patient cash price brought to you by Bristol Myers Squibb and Cost Plus Drugs”.
- No cost / markup / labour breakdown. The transparency module carries a single line: the manufacturer “will cover the markup, shipping, and dispensing fees for customers”.
- No “Retail Price at Other Pharmacies” and no “You save” — the module that produced this investigation’s adverse finding is not used here at all.
- Insurance exclusion disclosed prominently on the page: “Cost Plus Drugs does not currently accept insurance for this product.”
- The sponsorship is stated on the page, not buried.
Assessed on its own terms, the branded storefront is not hostile. It is accurately labelled, its price basis is stated, the insurance limitation is disclosed where a customer will see it, and the misleading savings module is absent. What it is not is the cost-plus-15 model, and a customer arriving from a page that itemises every dollar could reasonably assume otherwise.
One genuine disclosure gap sits inside it. The Terms carry manufacturer cash-programme clauses — including exclusions relating to government beneficiaries and true out-of-pocket credit — for a list of other manufacturers’ programmes hosted on the site. There is no equivalent clause for the Bristol Myers Squibb / Eliquis programme. Whether an Eliquis purchase here counts toward a Medicare Part D out-of-pocket total is therefore not disclosed anywhere on the consumer site, for the one branded product most likely to be bought by a Medicare-eligible customer.
The general principle this establishes for CHI: a structural alignment finding does not travel automatically across every product a platform sells. Cost Plus’s most powerful result is a property of one pricing architecture, and it should be cited as such — not as a general statement about everything on the site.
Section 10 · Evidence and method
How this was verified, and what remains untested.
The site is a JavaScript application that ordinary fetching cannot read, so the first-party record was captured through a rendering browser and read from the pricing payloads the pages themselves use. No account was created, no prescription was submitted, no purchase was made and no telephone call was placed.
The research package
12 research streams covering business model, price integrity, access, onboarding, fulfilment, customer service, refills and continuity, complaints, terms and policies, privacy, external validation and a dedicated contradiction search.
4 live first-party captures — the full FAQ with all 133 items expanded, the Terms, the Privacy Policy and the contact and product pages.
3 adversarial passes — a defence brief, a prosecution brief and a fact audit with a 132-row canonical facts table and a verbatim do-not-publish list.
The live verification pass
60 generic product pages drawn from the site’s 2,392-URL medications sitemap: 56 oral-solid pages at the default 30-count and 4 bottle-priced pages, spanning roughly 25 therapeutic collections from commodity statins to specialty oncology generics.
For each page the cost, markup, labour, shipping, calculated price, retail comparator and stock status were read from the site’s own data layer. Same-day controls were fetched independently for the identical strength, form and quantity, recording the displayed average retail price, the lowest non-membership coupon price at a named pharmacy, and the highest single pharmacy price listed.
Membership-tier prices, manufacturer savings cards and veterinary pharmacies were excluded from the controls.
An independent reality check
Two peer-reviewed studies bound how often Cost Plus is the cheapest option for a customer who already has coverage. Cost Plus was cheaper than the patient’s own out-of-pocket cost on 11.8% of all fills studied across payer types — private insurance 7.1%, Medicare 5.5%, uninsured 28.9%, Medicaid 0% — at a median saving of $4.96.
A later study of commercial generic fills found savings on roughly 15% overall, rising to nearly 80% where the patient was already paying more than $15, with median patient cost falling from $140 to $25 on fills above $100.
Both are the honest fact-check on “lowest prices anywhere”, and both show the model doing what it claims precisely where the customer is being hurt most.
Verification passes applied. A separate verification desk traced every number, date, quotation and URL in the dossier to a source file and found 19 defects — 4 requiring correction and 9 recommended — all of which were applied before the research was frozen. Zero do-not-publish breaches; zero arithmetic errors.
Named limitations, carried onto this page. No account was created, so the checkout screen, the auto-refill enrolment flow and the cancellation window remain untested. The live telephone test was not performed — the support finding rests entirely on first-party text, which is sufficient for what it claims and is not extended beyond that. Reddit was unreachable during the research window. The comparator’s third-party data vendor is unnamed by the company and was not independently identified. The acquisition-cost line is unaudited.
Confidence. Overall medium-high. Two components sit above that after the live pass — the price architecture (a 60-page audit verified in the data layer) and the comparator analysis (58 same-day matched controls) are both high confidence. The service layer and continuity findings remain medium, and the Terms-versus-FAQ operational reconciliation is unverified. The overall figure is held at medium-high by those gaps, not by doubt about the two central findings.
Final CHI interpretation
Cost Plus Drugs is the first company in this index whose central claim is a claim about mechanism rather than about service, and it is the first that can be checked to the cent on a public page. Checked that way, at catalogue scale, it holds. Sixty of sixty product pages reconcile exactly. The price is visible before the customer is. Eight of the ten hostile mechanisms catalogued in traditional pharmacy are removed or materially reduced, and every one that is removed is a rent mechanism — a way of taking money or data through a price the customer cannot see.
That is a real result, and it changes what CHI can say about every other company. Spread pricing, rebate retention, clawbacks, gag clauses, coupon-tiered pricing and discount-card data monetisation are not facts of the pharmaceutical industry. They are design decisions, and a company operating inside the same supply chain, buying from the same manufacturers, under the same regulator, has removed them and published the arithmetic. Ten named inverse mechanisms come out of this investigation as CHI’s first vocabulary for engineered non-hostility — not virtues, but page templates and fee schedules that operate on every transaction whether or not anyone is watching.
What Cost Plus has not engineered out is friction, and in places it has deliberately chosen it and said so. Below the product page it is an ordinary low-touch mail-order pharmacy: a form instead of a phone, a week instead of an hour, and the coordination work handed back to the patient. That trade is disclosed, and disclosure earns credit here. But the FAQ still tells customers to telephone a number the site does not publish, the contract still authorises what the FAQ promises not to do, and the savings module still measures against a benchmark the cash market abandoned years ago. Those are drafting failures inside an otherwise coherent structure — and each of them is fixable without spending a dollar more on service or accepting a cent less margin.
So the verdict is supports, not strongly supports, and the confidence is medium-high. Not because the alignment is doubtful — the two central findings are the best-evidenced in the investigation — but because a company that declares itself the customer’s ally is held to the standard it set, and on its own most prominent number it is still describing a saving the customer will not actually make.
The price it charges is exactly what it says.
The saving it advertises is measured against a world that no longer exists.
Reading this beside the rest of the index
Every other page here measures what a company takes. This one measures what a company removed.
Cost Plus is the control case. Read it against the companies at the friendly end of the index, and against the two where the same adverse pattern — a savings claim anchored to a price nobody pays — also appears.