Methodology
How JOYO collects brand context, generates marketing output, measures outcomes, and decides what to recommend next. This page is the required link target for any JOYO research, calculator, or comparison.
Brand context collection
JOYO ingests three sources: the business's website (homepage, product pages, about, pricing), social profiles (Instagram, LinkedIn, Facebook), and prior campaign assets when supplied. Extracted signals include voice (formal vs casual, sentence length, jargon density), offer (price band, dream outcome, time to value, effort), audience (demographics, jobs-to-be-done, objections), and proof (testimonials, case studies, named customers).
Generation rules
Every asset goes through three layers before output: (1) brand-fit check against the captured context, (2) platform-specific formatting per channel, (3) anti-AI-style filter that strips em-dashes, "delve", "leverage", "robust", "seamless", "embark", "navigate", and other LLM tell-tales. Outputs are first-draft quality and assume human review.
Measurement assumptions
Analytics inputs come from the user's connected ad accounts (Meta, Google) or manual uploads. JOYO computes channel-level ROAS and CPA, hook rate (3-second video view rate) for video creatives, and creative fatigue signals (frequency, CTR decay) per ad set. Calculations assume click-through attribution within a 7-day window unless the user changes the model.
Recommendation logic
JOYO ranks next-step actions by expected revenue impact (using historical ROAS as the prior), effort (rough hours estimate), and confidence (sample size, days of data). It surfaces the top three actions per week. Recommendations are never auto-applied; the human approves every change.
Research methodology
When JOYO publishes original research (e.g. SMB marketing time-waste survey), the methodology page links to: survey instrument, sample size, geographic coverage, time period, weighting (if any), incentive structure, and data download URLs. Sample sizes below n=100 are flagged with a small-sample caveat. Raw data is shared in CSV + JSON.
Comparison methodology
When JOYO publishes vs-comparisons (e.g. JOYO vs AdCreative.ai), every claim about a competitor cites the competitor's own pricing page, feature page, or G2 / Capterra listing with an "Accessed" date. Performance claims (CTR lifts, ROAS, time-to-launch) are made only with internal anonymized customer data, dated. Where JOYO lacks data, the comparison says so.
Update cadence
Methodology is reviewed quarterly. Material changes (new data sources, scoring model changes, voice rule changes) are versioned at the bottom of this page with date.
Contact
Questions about this methodology: hello@joyo.marketing.