This is the exact methodology behind every Riznefy engagement — how we audit, build, launch, and scale, and the 2026 platform-AI principles we build it on.
Most agencies sell a list of deliverables — ads managed, posts scheduled, reports sent. The Blueprint is different: it's the decision-making process underneath every account, the same four phases whether the budget is $2K or $200K a month. Tactics change by platform and season. This doesn't.
We don't skip phases to move faster. A system built out of order is the single most common reason paid media underperforms.
We map the current state of media, tracking, creative, offer, and conversion so the real bottleneck is obvious before anything else moves — not the bottleneck a vendor assumes because it's the service they sell.
Every audit starts with the numbers you already have: current ROAS or MER by channel, blended CAC, net margin, and where in the funnel drop-off is worst. Most accounts have one real constraint — a tracking gap, a weak offer, a creative that's fatigued, or a landing page that leaks — and the rest is noise until that constraint is fixed.
We also check Event Match Quality in Meta Events Manager and conversion tracking health in Google Ads and GA4 — a 2026 account with a sub-7 EMQ score is often "underperforming" for reasons that have nothing to do with creative or targeting at all.
Illustrative note: a full audit typically runs 5–7 business days depending on how many channels and how clean existing tracking already is.
Campaign structure, creative testing loops, landing-page priorities, and reporting standards get wired into one operating model — built once, reused every month, instead of rebuilt from scratch each time something breaks.
For paid social, this means architecture built for how Meta and TikTok's AI actually work in 2026 — consolidated ad sets instead of a dozen narrow ones, Advantage+ Audience on by default, and a creative production pipeline sized to feed the algorithm rather than starve it. For search and shopping, it means Smart Bidding structured around a real tCPA or tROAS, not a guess.
Reporting gets standardized here too — the same P&L-linked numbers (ROAS, MER, CPA, blended CAC) in every weekly check-in, so performance conversations stay anchored to revenue instead of drifting toward reach and engagement.
Illustrative note: build timelines vary by channel count — a single-platform build can be live in under a week; a full-funnel, multi-channel build usually needs the full two-to-three-week window.
We test in a controlled way, read the signal quickly, and remove the kind of complexity that makes scaling feel expensive — no big-bang launch, no fifteen ad sets fighting each other for the algorithm's attention.
Launch respects each platform's learning phase — Meta and TikTok typically need roughly 50 conversion events within a rolling window before delivery stabilizes, so budgets are set to reach that threshold within days, not weeks. Scaling only starts once ROAS has held steady for several consecutive days, and even then, budget moves in the 20–30% range every 48–72 hours rather than in one large jump that resets the learning phase.
Creative-as-targeting is treated as a first principle here: different hooks and angles are understood to reach different buyer segments on the same product, so "testing audiences" mostly gives way to testing creative variation instead.
Illustrative note: a ~50-conversion learning threshold and 48–72 hour scaling cadence are general 2026 platform guidelines — actual thresholds vary by ad account history and category.
Once the signal is stable, we widen what works, tighten what leaks, and keep growth anchored to margin instead of vanity — scale is a discipline, not just a bigger number in the budget field.
Winning ad sets get duplicated at 2–3x budget rather than having their original budget inflated directly — this resets delivery more gently and gives a cleaner read on whether performance holds at the new spend level. CBO campaigns are left to redistribute across ad sets rather than manually rebalanced, since fighting the algorithm's own allocation usually costs more than it saves.
Every scale decision is checked against the POAS target set in Phase 01, not against last month's number in isolation — a channel that's "growing" but sliding below target ROAS is a signal to tighten, not a reason to celebrate.
Illustrative note: a 2–3x duplication step and 20–30% incremental scaling range are general guidelines for 2026 platform delivery systems, not a guarantee of any specific outcome.
Every phase above runs on these principles — the assumptions that changed once Meta's Andromeda update and comparable platform-AI shifts made old-school manual targeting far less effective.
In the Andromeda era, the algorithm reads your creative and finds who responds — an "office worker with back pain" hook and an "athlete recovery" hook on the same product genuinely reach different buyers. We test creative angles the way accounts used to test audiences.
Advantage+ Audience runs on by default. Narrow interest-stacking is now a hint the algorithm mostly overrides anyway — we consolidate ad sets and let the platform's own signal-matching do the targeting work.
Every target ROAS is derived, not guessed: Target ROAS = 1 ÷ Net Profit Margin. A 35% margin sets a 2.86x floor — below that, spend is unprofitable no matter how good the surface-level numbers look.
Customer lists feed Advantage+ Audience and lookalike seeds. CAPI recovers a meaningful share of iOS-blocked events. Event Match Quality is treated as a KPI in its own right, not an afterthought in Events Manager.
On the strategy call, we walk through your actual account against this exact framework — live, not from a template.
