# OPPOSITION STRATEGY MEMORANDUM **Re: Proposed class certification — $7.95 priority-processing fee** --- ## 1. Bottom-line recommendation Combine, in this order of relief requested: 1. **Primary:** Deny certification **on the present record and without prejudice** — plaintiffs bear the burden and have offered one marketing page and a queue dataset against a record showing four materially different disclosure paths, no uniform representation, no damages model, and no representative for three of the four paths. 2. **Alternative:** If the court is inclined to entertain any narrower class, **defer briefing 45–60 days and compel seven identified categories** of named-plaintiff documents that plaintiffs agreed to produce by June 1. 3. **Not recommended:** a broad six-month stay (Rule 8 forbids unfocused delay and asking for it will forfeit the credible narrow request), or full merits briefing now (invites a "you're conflating merits" ruling and commits us before our own record is built). These two positions are consistent, and the brief must say why: **we do not need discovery to defeat plaintiffs' motion as filed; we need it before the court could certify anything, and to prevent plaintiffs from curing a documentary gap with new declarations about records they are withholding.** --- ## 2. The three representatives on the present record **Milo Chen (mobile).** Probably has standing if a net $7.95 debit is shown; that is not yet in evidence. His testimonial gaps (whether the box was preselected, whether he opened the linked free-option explanation) matter less than plaintiffs fear and less than we might wish: if our own logs place him in the 21,300-member mobile cohort, the interface was standardized, and Rule 5 permits inferring reliance from a materially uniform representation. **Honest assessment: Chen is the plaintiffs' best asset and a plausible representative for a mobile-only subclass.** He is not typical of the desktop cohort (affirmative unchecked box), the legacy cohort (fee only in linked terms), or the call cohort. His failure to recall opening the link is relevant to the contract claim (assent to a disclosed term) but does not defeat typicality within a uniform-screen subclass. **Tara Vale (calls).** The weakest. Her claim rests on an alleged oral statement she cannot date or attribute, in a cohort exposed to **six** different scripts. She cannot establish which script she heard, so she cannot show her exposure is typical of anyone, and her claim requires witness-by-witness proof — the paradigm of individualized liability. **Caution:** we retained only sample recordings. Arguing "she cannot prove what she was told" invites a retention/adverse-inference response. The correct framing is not that plaintiffs lack proof but that **oral variation across six scripts makes common proof unavailable to either side**, which defeats predominance for the 9,200-member call cohort regardless of who bears the gap. **Jae Morgan (refund status).** The cleanest defect. Under Rule 4 an unreimbursed economic loss is an element; if "CB-R" means a completed chargeback, Morgan has no injury and no standing, and cannot represent anyone. **But the ambiguity is in our ledger**, so we must not rest on Morgan's silence: reconcile the code against processor settlement records ourselves and present that reconciliation. Morgan's card statement resolves the question conclusively in one page. Morgan's enrollment path is also unstated in the record — a second, independent typicality gap. **Adequacy.** No conflict-of-interest problem is apparent. The failed June 1 production is better used as the predicate for Rule 8 relief and for a preclusion order than as a standalone adequacy attack, which courts discount. --- ## 3. Common questions vs. path-dependent questions **Genuinely common (concede these; Rule 9 credibility depends on it):** - Numerosity (62,400). - The fee's existence, amount, and uniform billing mechanics. - Whether the shared backend queue in fact delivered differentiated processing — the 18-hour weekday improvement and the absence of weekend improvement is common evidence about a single product. - StreamLedger's uniform internal decision to offer and price the fee; its refund policy. **Path-dependent (each tied to an element, per Rule 9):** | Variation | Element it defeats | |---|---| | Desktop unchecked box disclosing the fee **and** the free alternative | Exposure to a deceptive representation/omission; also Rule 5 bar on recovering a fee "knowingly and separately selected" | | Mobile preselected box with linked free option | Materiality; whether preselection plus a link is deceptive — uniform within cohort, different from all others | | Legacy landing page, fee only in linked terms | Omission theory and incorporation-by-reference assent — a distinct legal question | | Six call scripts, sample recordings only | Exposure and causation; no uniform representation, so no inferred reliance | | Weekday vs. weekend use | Causation and the "no benefit" premise; benefit was real for weekday users | | Differing assent artifacts and terms versions across 36 months | Contract claim's core: "terms presented and assented to" | | Refund / chargeback / employer-card payment | Unreimbursed economic loss; class definition sweeps in uninjured members | The complaint's operative allegation — that the fee was **"uniformly hidden"** and provided **no benefit** — is contradicted by StreamLedger's own records on both halves. That is the spine of the opposition. --- ## 4. Predominance **Consumer statute.** Rule 5 allows inferred reliance only from a **materially uniform** representation. There is none class-wide: one cohort saw an express disclosure plus the free alternative; one saw a preselected box; one saw nothing outside linked terms; one heard six scripts. Plaintiffs' single marketing page cannot bridge this — and no named plaintiff has testified to seeing it (Chen was on mobile). Exposure, causation, and unreimbursed loss would each require path-level or individual proof. Predominance fails as proposed. **Contract.** The claim is defined by the terms presented and assented to, so it is path-specific by construction, and further fragmented by interface and terms-version changes over three years. For the desktop cohort, Rule 5 affirmatively bars recovery of a separately selected disclosed fee. There is no coherent class-wide breach theory. **Superiority.** Aggregate exposure is ~$496,000 gross, ~$444,000 net of the 5,400 refunds and 1,180 chargebacks (subject to overlap). Argue that a **claims-administered refund program**, which StreamLedger has already run for 5,400 customers, is a feasible superior alternative to a proceeding requiring four subclasses, script-by-script testimony, and 7,900 individual reimbursement inquiries. Do not overclaim: for small-value claims courts often find class treatment superior, so pair this with the narrow fallback in § 6. --- ## 5. Damages model (Rule 6) Plaintiffs have offered **no model**. The one implied by the complaint — a full $7.95 refund to all 62,400 — fails Rule 6 twice over: - **Theory mismatch.** A worthlessness/full-refund model contradicts the 18-hour weekday improvement. If the theory is nondisclosure, the measure is the price of an unwanted option, which may require a price-premium or conjoint analysis calibrated to each path's disclosure — none exists. - **Mandatory deductions.** 5,400 full refunds and 1,180 chargebacks must be removed (with de-duplication; some customers may appear in both). The 7,900 employer-card payments cannot simply be deducted or included: the field establishes only the instrument, not who bore the loss. Because unreimbursed loss is an **element**, not merely quantum, this is a standing-and-scope defect, not a damages detail. Present it as requiring either exclusion or a screening mechanism — do not claim it independently defeats predominance, since individual damages computations often do not. Also note the **average** conceals distribution: an 18-hour weekday mean means some weekday users got little or no improvement. Present the distribution honestly; it shows benefit varied by customer and date — which defeats *both* a uniform-benefit and a uniform-no-benefit theory and confirms individualized causation. --- ## 6. Issue class and subclasses (Rule 10) — the fallback we can live with Do not tell the court "nothing is certifiable." Offer the narrow structure: - **Certifiable as an issue class:** whether the shared queue delivered differentiated processing, separately determined for weekday and weekend submissions. Pure common fact, cleanly severable. Note candidly that it resolves no element by itself, so its practical value is limited. - **Potentially certifiable subclasses:** mobile (21,300) and legacy (17,100) — each has a standardized presentation supporting a common question. - **Not certifiable:** the desktop cohort (Rule 5 bar; no deception on the face of the disclosure) and the call cohort (six scripts, oral variation, sample-only recordings). - **Decisive point now:** even the viable subclasses **lack representatives**. Chen is mobile-only; Vale is call-only; Morgan's path and injury are both unestablished. No subclass can be certified today for want of an adequate representative — which is precisely why the court should deny without prejudice rather than certify a restructured class on this record. --- ## 7. The named-plaintiff materials that matter — and exactly why Present this as a chart, with **decisional significance stated as a question each document answers, not as a predicted answer** (Rule 8 requires the former; candor forbids the latter). 1. **Confirmation emails / enrollment screenshots** — fixes which of the four paths each plaintiff used and what that plaintiff actually received, including the preselection state and the free-option link. Goes to standing (exposure), typicality (subclass fit), predominance (uniformity). Our logs show the path; the plaintiff's email shows the *rendered version* the plaintiff received. 2. **Bank/card statements** — whether $7.95 was in fact debited and whether any credit reversed it. Goes to injury for all three and resolves Morgan's "CB-R" definitively. 3. **Refund and chargeback records/correspondence** — completed vs. rejected chargeback; whether a refund was accepted. Element of unreimbursed loss; class-definition overbreadth. 4. **Employer-reimbursement records** (expense reports, reimbursement receipts) — whether the plaintiff personally bore the fee. Element of loss; representative of the 7,900-payment problem. 5. **Enrollment communications** (emails, chats, notes, call confirmations) — for Vale, anything fixing date, number, or representative so the call can be mapped to one of six scripts and any retained sample recording located. 6. **Terms version received** (legacy path) — what was incorporated and how; contract assent. 7. **Submission-date records** — whether the plaintiff filed on weekdays or weekends. Causation and the "no benefit" premise. State plainly: these documents may corroborate plaintiffs' accounts. The point is that Rule 7 requires certification to rest on **evidence**, and the movant controls the evidence. --- ## 8. Twenty-one-day filing sequence **Days 1–2 (internal).** Litigation hold; preserve and version-stamp all four interfaces; pull archived screenshots by version; reconcile Morgan's "CB-R" against processor settlement data; de-duplicate refunds and chargebacks; run the queue dataset as a weekday/weekend distribution, not just means; inventory the six scripts and the sample-recording coverage by period. **Day 3.** Deficiency letter to plaintiffs: the seven categories, the June 1 agreement, decisional significance, demand for production in five days, and an offer to consent to a short reset of the opposition deadline in exchange. **Days 4–6.** Conference call; document the refusal. **Day 7.** File **Motion to Defer Certification Briefing and to Compel Targeted Production**: seven enumerated categories, a decisional-significance chart tied to standing/typicality/predominance, a 45-day schedule (not six months), a request to reset the opposition deadline, and a request for expedited hearing. Include a preclusion request: plaintiffs may not support certification with declarations about matters covered by withheld documents. Ask for fee-shifting on the broken agreement. **Days 7–15.** Build the opposition record regardless of the motion's outcome — five declarations: product/UX (four paths by version, with exhibits), engineering (queue performance distribution, weekday/weekend), payments (refund/chargeback reconciliation, de-duplication, the limits of the business-account field), call center (six scripts, retention practice), and records custodian (the "CB-R" ambiguity and what the processor data shows). Serve targeted evidentiary objections to plaintiffs' marketing-page exhibit (authentication; no showing of exposure by any named plaintiff). **Days 16–20.** Draft and finalize the opposition as a **freestanding** document that does not depend on the deferral motion being granted. Structure: (i) concede numerosity and identified common issues; (ii) typicality/adequacy/standing per plaintiff, with Morgan's injury front-loaded; (iii) the four-path variation matrix tied to statutory elements; (iv) contract claim's assent dependency; (v) no damages model, mandatory deductions, overbroad class definition; (vi) superiority and the refund-program alternative; (vii) fallback: if anything is certified, only a weekday/weekend-split issue class or mobile/legacy subclasses, and only with representatives who exist. **Day 21.** File the opposition, expressly preserving the deferral request and reserving a post-production standing challenge to Morgan and further limited depositions of all three plaintiffs. **One discipline throughout:** never argue that "every customer is different." Argue that *this* variation defeats *that* element, and say only what each missing document would resolve — in either direction.