8 cases, 8 brands, one methodology.

The book of business across 8 cases reduces to roughly 8 distinct brands. That's a real client list. It just doesn't look broader by being split into 8 cards. Consolidating to 5 cases (one per signature client + one anonymous methodology case) would read as more rigor, not less proof.

• Aspirational claims stand in for evidence. "We don't just drive awareness — we drive tangible sales and market leadership." Repeated across cases 2, 4, 6 in different forms. Not findings, just affirmations.
• Generic strategy vocabulary repeats. "Data-driven audience segmentation," "real-time optimization," "adaptive targeting," "dynamic optimization" appear in nearly every case with no description of the actual data, models, or decision rules.
• No tradeoffs disclosed. No case mentions what didn't work, what was reallocated mid-campaign, what budget shifted from where, what the team learned.
• No team or process detail. No mention of who staffed the engagement, what the planning cadence was, which measurement partner validated the read, which weeks shifted strategy.
• Channel lists read as menus. Case 2's channel list spans 13 properties (NFL, MLB, Shark Week, Olympics, MNF, debates, college football, Apple TV Plus MLB, Howard Stern, REELZ, The Voice, digital). The reader cannot tell which channel drove which result. Each channel should ladder to one outcome.

1. Consolidate to 5 cases. One per signature client (Flex Seal, Pooph, Razor, Zevo, TIDL) plus the anonymous methodology case. Cuts overlap, deepens each piece, signals editorial discipline.
2. Add a "How we measured" line to every case. One sentence under every Results block: "Sales data via Circana scan, period vs. prior 4-week baseline, no held-out control" is more credible than no methodology.
3. Standardize the time window per case. Pick one window (e.g., first 8 weeks) and report all metrics against it. Stop mixing "first week," "first month," "month-over-month" inside the same Results block.
4. Add caveats to the five biggest numbers. +400%, +352%, +342%, +59%, 8:1 ROAS. Each gets one qualifying phrase. The numbers become more credible, not less.
5. Add one attributed client quote per case. Eight quotes across the existing client relationships. Cheapest credibility unlock on the site.
6. Map every channel in the list to one outcome. Case 2's 13-property channel list becomes: "MNF drove the household reach, REELZ drove the late-night frequency, Howard Stern drove search lift week 3." Each channel ladders to a job.
7. Show one campaign artifact per case. A creative still, a planning grid, a Tuesday read screenshot. The portfolio currently has zero visual evidence of work.
8. Remove "<$5 CPM" buying brags without context. Either show the benchmark and the daypart/inventory tier, or remove. Reads as media-buyer chest-thumping in a CMO-facing piece.
9. Name the anonymous case. If the Zip Code Targeting case is the most methodologically rigorous, naming the brand strengthens it most. If the client truly won't let their name be used, say "national CPG, brand anonymized at request" to make the anonymity intentional rather than missing.
10. Add a Methodology / Measurement page. One-page summary of how Diray measures lift across the standard cases: test cells, control design, lift study partners, MMM cadence, brand-lift survey vendor. The page becomes the link every case-study Results block points to for "How we measure."