How PrismSpy tracks Indian healthcare Meta ads: data sources, refresh cadence, scoring

Every time we onboard a new agency client to PrismSpy, the first question is some version of "where does the data come from?" It is a fair question. Competitive intelligence tools have a reputation for opaque sourcing, dubious scraping, and vague accuracy claims. PrismSpy runs on public disclosure sources only, refreshes on a documented cadence, and scores creative on healthcare-specific criteria that we can walk anyone through. This guide is that walkthrough.

The design philosophy is simple. Everything PrismSpy shows can, in principle, be assembled by a healthcare marketing team using free public tools. We built the platform because doing it by hand at the scale needed for a healthcare agency running 10+ accounts is a full-time job — and the interpretation layer on top matters more when the data is clean and pre-clustered.

Data source 1: Meta Ad Library

The primary source. Meta publishes an Ad Library that discloses every currently active ad across Facebook, Instagram, Reels, Messenger, and Audience Network. This is a public transparency surface Meta maintains globally. PrismSpy consumes this feed for every tracked Indian healthcare brand on our curated watchlist — currently 250+ brands across all major specialties.

What we capture per ad: creative (image, video URL, carousel cards), ad copy (headline and body), advertiser page name and ID, first-seen date (when the ad went live), active platforms, active countries of delivery, and the landing URL. What we do not capture (and could not, from public sources): actual ad spend, targeting, performance data (CTR, CPC, CVR), or historical ads that have stopped running (for non-political categories).

Data source 2: Google Ads Transparency Centre

Google runs its own separate transparency surface at adstransparency.google.com. It covers Google Search, YouTube, Display, and Discovery ads. PrismSpy consumes this feed alongside Meta because Indian healthcare buyers advertise on both platforms and understanding one without the other leaves half the picture.

Google Transparency exposes slightly different fields than Meta. You get creative, format, first-seen date, and the advertiser name, but Google does not always disclose active countries at the same granularity. We cross-reference with Meta pages to confirm the same brand entity when there is ambiguity.

Data source 3: public landing pages

Every ad has a click destination. PrismSpy captures a snapshot of that landing page — HTML, hero copy, form fields, price mentions — on the day the ad is first seen. When the landing page changes materially (offer copy update, form field change, hero copy shift), we snapshot again and diff.

This gives us two things: current-state landing page intelligence (what is the offer on the page right now) and temporal intelligence (how the page has evolved across the ad's lifecycle). The temporal view is useful — if a competitor's landing page changed twice in the past month, that is a signal they are still iterating and possibly not fully converted yet.

What we deliberately do not do

Three things worth stating explicitly:

• No scraping of authenticated dashboards. We do not log into ad accounts, agency dashboards, or private surfaces. Everything is public disclosure.
• No CRM or lead data. PrismSpy does not touch client lead flow or performance data unless the agency separately connects those systems for internal analysis (that is a different product).
• No private ad performance data. If it is not on Meta's public library or Google's public transparency, we do not have it. Anyone selling you "real competitor CPCs" is either fabricating or scraping something they should not be.

Refresh cadence

PrismSpy runs a nightly full snapshot for every tracked brand. That gives 24-hour resolution on ad launches, pauses, offer changes, and creative refreshes. For high-priority brands (aggressive spenders in the current period, brands in launch phase, brands with unusual velocity), we run twice-daily snapshots to catch the pause windows that would otherwise take 24 hours to reflect.

The dashboard shows the freshness timestamp on every panel. If a brand's last snapshot was 8 hours ago, the panel shows that. This is deliberate — competitive intelligence loses value fast when it is stale, and hiding the freshness timestamp is one of the most common failure modes in horizontal tools.

Specialty clustering

Brands are pre-assigned to specialty clusters based on their historical ad portfolio. Current clusters cover IVF, dermatology, hair transplant, aesthetic, dental, hospital (multi-specialty), ophthalmology, cardiology, orthopedic, oncology, gastro, pediatrics, and a few smaller specialties. Multi-specialty hospitals appear in multiple clusters based on ad-copy signals.

Within a brand, ads are further sub-clustered by treatment focus using ad copy and landing page text. This is what lets the dental cluster split ads by aligners vs. implants vs. smile makeover automatically instead of requiring a human to tag each one.

Creative scoring: calibrated on the vertical

PrismSpy scores every ad on a composite creative quality score. The composite is built from three inputs:

1. Run length. Longer is better. Ads still running at 90+ days score materially higher than ads that turned over quickly.
2. Format-intent match. An ad whose format fits the underlying buying decision (video for procedural education, carousel for before-after, static for price anchor) scores higher than a format-treatment mismatch.
3. Structured criteria. Hook clarity (specific vs. vague), proof presence (doctor named, patient count, before-after present), offer specificity (concrete vs. abstract).

The scoring is calibrated on Indian healthcare specifically. General "creative quality" scoring from horizontal tools does not translate — a hook that reads as vague in beauty ads may read as appropriately restrained in healthcare, and vice versa. Every specialty cluster has slightly different weighting on the structured criteria.

Compliance signalling

PrismSpy flags ads that appear to breach four specific compliance regimes: NMC guidelines (doctor self-promotion, celebrity endorsement, claims substantiation), ART Act 2021 (IVF-specific claims), PCPNDT (sex-selection language), and ASCI (substantiation of clinical claims). Flags are surfaced in the ad detail view with the specific rule that appears to be triggered.

These are algorithmic flags, not legal opinions — a flag says "this ad has patterns that historically trigger complaints under this regime", which is different from saying "this ad is non-compliant". Legal review sits with your compliance team. The flags are useful for two things: understanding what the market gets away with in practice, and knowing what your own creative team should avoid.

What PrismSpy is not designed for

Non-Indian markets. The platform architecture works globally but the curated brand watchlists, offer benchmarks, and compliance signalling are calibrated for India. Most PrismSpy value comes from the Indian healthcare depth, not the raw data.

Non-healthcare verticals. Aesthetic clinics, dental clinics, hospitals, IVF, dermatology, and adjacent specialties are covered. Consumer beauty brands that are not clinic-owned, wellness apps, fitness brands, and telemedicine are outside the scope. If you need horizontal breadth, use SocialPeta or AdScout instead — they are built for that.

Real-time bid data or performance data. Public transparency does not expose it. We do not fabricate it. The Meta Catalyst IQ product handles the campaign-launch and performance-attribution work downstream, using your own connected ad account data.

Frequently asked