CRED Social Intelligence Platform

Process Documentation & Pricing Guide

Including: Celebrity Affinity

Version: 1.0 | Classification: Internal Wiki | Last Updated: February 2026

Overview

CRED Social Intelligence is an automated data pipeline that resolves celebrity or athlete names to verified social profiles across multiple platforms, scrapes follower and engagement metrics, and delivers clean, normalised reports in the format a client requires. The only required input is a list of full names.

Give us a name — we return complete social media profile metrics and post engagement data across every major platform. No manual research. No guesswork.

How to Submit a Request

All data requests are submitted through the CRED AI Assistant. Clients interact with the assistant directly to define their universe, select platforms, specify a target geography, and choose a report format. The assistant handles job creation, status updates, and delivery.

Typical submission flow:

1. Open the CRED AI Assistant

2. Provide a list of names and your brief (geography, sector, platforms)

3. The assistant confirms scope and credit cost before running

4. Data is returned in your preferred format within the agreed SLA

No technical knowledge is required. The assistant translates a plain-language brief into a structured data job.

Pipeline — Step by Step

Step 1 — Client Submits Names

The client provides a list of celebrity or athlete full names via the CRED AI Assistant. No social handles, profile URLs, or platform-specific identifiers are required.

Example input:

"Harry Kane"  |  "Jamal Musiala"  |  "Alexander Zverev"  |  "Dirk Nowitzki"

Step 2 — Automated Profile Discovery

An automated web search resolves each name to a verified, official social media profile. The resolution engine handles name ambiguity, verified account badges, and cross-platform matching to ensure accuracy before any scraping begins.

• Name resolution via automated web search

• Verified badge cross-referencing

• URL verification before dispatch

• Platform routing — each URL classified to the correct scraper

Step 3 — Platform-Specific Scrapers

Each verified URL is dispatched to a dedicated scraper built specifically for that platform's data structure. Scrapers run in parallel across all seven supported platforms.

Step 4 — Two-Layer Data Extraction

Each profile is scraped across two dimensions simultaneously.

Step 5 — Cleaning & Normalisation

Raw data from seven different platform structures is unified into a single consistent schema. This step includes deduplication, validation, and QA checks before delivery.

• Unified schema across all platforms

• Deduplication of cross-platform profiles

• Engagement rate calculation: (avg likes + avg comments) / total followers × 100

• Note: engagement rates above 100% are valid — they indicate posts reaching beyond the follower base through algorithmic amplification

Step 6 — Report Delivery

One row per athlete per platform. Every metric, every platform, structured and ready for analysis. Delivered via the CRED AI Assistant in the format the client requests within the agreed SLA.

• Output formats: CSV, XLSX, JSON, Google Sheets

• Optional: formatted narrative report with ranked tables, tier classifications, and athlete insights

Report Types

Reports follow CRED's standard templates for charts and structured outputs. All report types are generated automatically from the same underlying data pull and delivered via the CRED AI Assistant.

Raw Data Export

The standard output: one row per athlete per platform. Suitable for clients who want to perform their own analysis or feed data into internal tools.

Formatted Intelligence Report

A structured, client-ready report built on CRED's standard report template. Includes ranked tables, tier classifications, engagement analysis, and sponsorship fit assessments — ready for direct insertion into client pitch decks.

• Ranked tables by audience size, engagement rate, or combined following

• Tier classification based on market concentration

• Top athlete profiles with sponsorship insight commentary

• Methodology notes and data confidence flags (Actual vs. Estimated)

Market Intelligence Report

A specialised report focused on audience reach within a specific target market or geography. Uses actual scraped data where available, supplemented by Celebrity Affinity (see below) for the full universe. Output follows CRED's standard chart and report template, with all figures clearly flagged as actual or estimated.

Celebrity Affinity

Celebrity Affinity is a feature of the CRED Social Intelligence Platform that estimates how much of a personality's social audience is concentrated in any target market. It is market-agnostic, works across all supported platforms, and is embedded directly into CRED's standard report templates and chart outputs.

Background & Rationale

Per-country audience data (the percentage of a profile's followers from a specific country) is not publicly accessible on any social platform. It exists only behind the platform's own analytics paywall — available exclusively to the account owner or their authorised representative.

Celebrity Affinity provides a calibrated best-estimate of a personality's audience size within any target market, using a validated anchor dataset as the baseline. The model can be applied to any geography by updating the anchor parameters — no custom development required.

No third-party vendor or data provider has access to real per-country follower counts at scale. Celebrity Affinity is the best available estimate outside of a direct data-sharing agreement with the platforms themselves.

How Celebrity Affinity Works

Celebrity Affinity requires one anchor data point: a rights-holder (e.g., a club, league, or brand) whose geographic audience breakdown is known from a direct CRED data pull. From that single anchor, it estimates any individual personality's audience in the same target market using two independent methods, then returns the average as the primary figure.

Variable Reference

V1 — Conservative Estimate

V1 models the personality's market audience as their proportional share of the anchor's known market audience, scaled back to their own follower base. Smaller accounts with less market exposure receive a lower estimate.

V1 = F × (F / G_anchor)
V1 = F² / G_anchor

Example (Germany, anchor: Bayern Munich — G_anchor = 96.29M, P_anchor = 0.62)

Harry Kane, 18.07M Instagram followers:

V1 = 18,068,051 × (18,068,051 / 96,290,000)
V1 = 18,068,051 × 0.1877 = 3,390,326

V2 — Optimistic Estimate

V2 applies the anchor's market concentration flat to every personality, assuming their follower base has the same geographic mix as the anchor rights-holder. This produces an upper-bound figure.

V2 = F × P_anchor

Example (same parameters)

Harry Kane:

V2 = 18,068,051 × 0.62 = 11,202,192

Average — Final Reported Figure

The average of V1 and V2 is used as the primary reported market audience figure in all client-facing outputs.

Avg = (V1 + V2) / 2

Example:

Avg = (3,390,326 + 11,202,192) / 2 = 7,296,259  (~40.4% German audience share)

The Germany / Bayern Munich example above is illustrative. Celebrity Affinity applies the same formula to any target market — swap in the relevant anchor rights-holder's data to recalibrate.

Calibration Adjustments for Non-Anchor Personalities

For personalities outside the anchor dataset, Celebrity Affinity applies additional adjustments based on nationality, club market, league, and content language.

Percentages above are illustrative for the Germany example. Recalibrate for each new market using the relevant anchor dataset.

Data Confidence Flags

All reports flag each personality's market audience figure as either Actual or Estimated.

• Actual: Sourced from two independent CRED platform scrapes (v1 + v2), with the median as the primary figure.

• Estimated: Derived from Celebrity Affinity with calibration adjustments applied.

Pricing

Credit Model

CRED Social Intelligence operates on a credit-based system. Credits are consumed per personality per data pull, based on the number of platforms included.

Default scrape covers Instagram and X/Twitter. Additional platforms can be added to any request. A full five-platform pull for a universe of ~40 personalities sits comfortably under 100 credits.

Credit Consumption

Report Add-On

Raw data export is included in the base credit cost. Formatted reports are available as an add-on.

Pricing Notes

• Credits are calculated per personality, not per platform — multi-platform requests are bundled.

• Pricing is based on a universe of approximately 40–50 personalities. Larger universes (100+) should be discussed with the CRED team for volume pricing.

• Re-runs of the same list (e.g., quarterly monitoring) consume the same credits as the original pull.

• Celebrity Affinity estimates do not consume additional credits — they are generated at report compilation time from the anchor dataset.

• Validated actual market audience data for non-anchor personalities requires a dedicated geographic pull — contact CRED for pricing.

Known Limitations & Data Notes

Geographic audience data: Real per-country follower counts are not publicly available on any platform. Actual v1/v2 scrape data applies to anchor dataset personalities only. All other figures are Celebrity Affinity estimates. No competitor or vendor has access to this data without a direct API agreement with the platforms.

Engagement rate calculation: Calculated as (avg likes + avg comments per post) / total followers × 100 from a scraped sample of 12–24 recent posts. Rates above 100% are valid and indicate algorithmic amplification beyond the follower base.

Profile discovery gaps: Personalities with no verified public social presence, non-standard name formats, or very low social footprints may not be resolved. These are flagged as "not found" and excluded from ranked outputs.

TikTok: Available as part of Standard+ or Full Platform pulls. Any TikTok figures included without a dedicated pull are directional estimates based on cross-platform benchmarks.

Data freshness: All data reflects the state of each profile at the time of collection. CRED recommends re-running scrapes quarterly or ahead of major campaign decisions.

Recommended Workflow

1. Submit your brief to the CRED AI Assistant — names, target market, platforms, and preferred output format.

2. Review the scope and credit cost confirmation before the job runs.

3. Use the anchor dataset table (actual data) for high-confidence recommendations; use the full universe table for broader market briefs and pitch decks.

4. Filter by Tier for use-case targeting: Tier 1 for market-focused campaigns, Tier 2 for cross-market plays, Global for awareness-first strategies.

5. Flag predicted market figures clearly in client deliverables. Validate with an actual v1/v2 CRED scrape before final recommendations if budget allows.

6. Revisit quarterly — audience data shifts with transfers, tournament performance, and viral moments.

CRED Investments — Confidential — Internal Wiki Documentation