News · Product
Sponsorlytix launches sponsorship dashboard and tracking portal
Motivation and problem formulation
Sponsorship value is usually argued from rate cards, GRP-style proxies, or post-hoc agency audits. Those methods struggle when inventory is fragmented across live broadcast, delayed rights windows, club and fan social, news coverage, and esports streams — and when a single brand appears as kit marks, LED ribbon, static boards, digital overlays, and organic mentions in the same week.
We formulate the task as multi-asset sponsorship measurement: given a rights holder’s season or tournament corpus, recover (i) which brands appear, (ii) on which asset classes and placements, (iii) under what visibility conditions (position, time-on-screen, clarity, size, viewership context), and (iv) how those observations combine into a defensible brand-value estimate. The decision users care about is commercial: renew, reprice, reallocate inventory, or underwrite a new deal.
Honest constraints at launch: we did not publish a public API, a fixed public list of every asset taxonomy node, or a peer-reviewed benchmark leaderboard. What we shipped was an operational measurement stack with client dashboards, a tracking portal, and case-study reports that demonstrate end-to-end coverage on named competitions.
Architecture schematic (qualitative) — not a published latency benchmark.
System architecture
Sponsorlytix combines computer vision (logo and placement detection on broadcast and visual social), language models (news and caption context), and proprietary valuation factors — position, time, clarity, visibility, size, and viewership. Detection outputs feed placement analytics; analytics feed live dashboards and customized reports; pre-invest ROI tooling (later productized more explicitly as Estimator) sits on the same measurement family.
At launch the primary surfaces were: (1) a sponsorship dashboard for real-time and historical views, and (2) a sports detection & tracking portal for operational review of detections. Intended audiences include clubs (pro and semi-pro), federations/leagues, brands, athletes, esports organisations, and media/rights holders.
Illustrative evaluation profile — not a published benchmark.
Data and methods
Season- and tournament-scale corpora illustrate how the system runs in practice. For example, the Chelsea FC Season 2025-26 Data Insights report states that Sponsorlytix used in-house AI models with ML, CV, and LLM components and that “all games full broadcast, social media posts, news articles, stadium fans and much more were tracked, detected and measured.” The earlier WTA Makarska Open 2022 analysis (published July 2024) notes live broadcast on beIN Sports (MENA) and HRT (Croatia), with 13 additional delayed broadcasters, and describes measurement across social interactions, broadcast exposure, naming rights, and charity/community signals. That WTA write-up also references an eleven-asset breakdown — consistent with the multi-asset coverage we design studies around.
Evaluation protocol. For full-season or full-tournament studies we treat the unit of analysis as the rights period. Primary observables are brand detections, placement class, and factorized exposure. Downstream outputs are brand-value estimates and commercial breakdowns (tiering, revenue-context charts). We do not publish raw mAP, frame-level precision/recall, or latency SLAs as open benchmarks in this article. Where numbers appear below, they come from published case PDFs under media/resources/, with captions that name the source.
From Chelsea FC Season 2025-26 Data Insights (Sponsorlytix case study).
Results from published case materials
The Chelsea season report further breaks club-side revenue context for 2025-26 (TV rights $252M, sponsorship $175M, ticketing & hospitality $118M, merchandising $63M, licensing $15M, stadium tours $7M; costs shown at −$720M) and reports cumulative TV audience figures with regional breakdowns (e.g. MENA 202M, Africa 115M, Asia 109M in the territory charts). Those figures are commercial context for sponsorship measurement, not detection accuracy scores.
The FIFA World Cup 2026 Case Study scopes 104 matches across 16 host cities (11 US, 3 Mexico, 2 Canada), 48 teams, and a 39-day window (11 June – 19 July 2026). The same PDF reports projected host economic impact exceeding $5 billion and 40,000+ roles created — again economic context surrounding the sponsorship measurement exercise, not model accuracy.
From Chelsea FC Season 2025-26 Data Insights (Sponsorlytix case study).
Capability comparison (suite context)
| Capability | Sponsorlytix (core) | Estimator | Dantani Sports |
|---|---|---|---|
| Logo / placement CV on broadcast & social | Primary | Consumes measurement | Not primary |
| Live sponsorship dashboards | Primary | Adjacent | Media dashboards |
| Deal valuation / pre-invest ROI workspace | Tooling lineage | Primary | AVE / media value |
| Multilingual news + social listening | Partial (news assets) | — | Primary (160+ languages claimed) |
| Auditable commercial artifacts | Customized reports | Explicit design goal | Partner reports |
Qualitative capability matrix across the suite — not a scored bake-off.
Evaluation table (metrics we measure; open results withheld)
| Metric | Definition | Unit | Protocol | Result |
|---|---|---|---|---|
| Brand detection quality | Correct brand ID given visible mark | Precision / recall | Held-out annotated frames per competition | Internal eval — contact sales |
| Placement classification | Correct asset/placement class | Accuracy | Labelled placement taxonomy | Internal eval — contact sales |
| Time-on-screen estimate | Duration of qualifying exposure | Seconds | Broadcast timeline reconciliation | — |
| Brand-value reproducibility | Variance of value under factor ablation | % / $ | Factor leave-one-out on case corpora | Internal eval — contact sales |
| Report latency | End-to-end study turnaround | Hours / days | Operational SLAs by package | — |
Limitations and failure modes
- Occlusion and motion blur on pitch-side LEDs and kit marks during fast camera pans reduce detection confidence; we treat low-clarity frames as down-weighted rather than forced positives.
- Lookalike marks and co-branded kits create identity ambiguity; human review remains part of high-stakes commercial deliveries.
- Social organic vs paid distinctions are not always recoverable from public posts alone.
- Viewership multipliers depend on rights and panel sources that vary by territory; brand-value is sensitive to those inputs.
- Open asset taxonomy: we keep the multi-asset inventory list illustrative rather than locking a closed “10 assets” enumeration in this post; Dashboard 2.0 (mid-2026 rewrite) feature delta vs the launch dashboard continues to evolve.
- Image-only case PDFs (e.g. Al Ain FC, BKFC Dubai) are cited by title here; numeric extraction from those scans is out of scope for this article.
Safety and responsible use
Sponsorship measurement can affect commercial negotiations and athlete/club narratives. We design for auditable artifacts, avoid presenting internal detector scores as universal accuracy claims, and expect clients to keep human commercial judgment in the loop. Brand detection on fan content raises privacy and platform-policy considerations; deployments should follow applicable rights, consent, and platform terms.
What’s next / availability
Sponsorlytix remains live at sponsorlytix.ai with Client Login and demo CTAs. Estimator is available as the valuation workspace in the same product family (announced 30 June 2026). Further case studies (Chelsea 2025/26, FIFA World Cup 2026, WTA Makarska, and others listed under Resources) illustrate coverage at season and tournament scale.
Appendix — glossary
Brand value (Sponsorlytix sense). Proprietary combination of exposure factors into a monetary estimate for a sponsor over a rights period.
Placement. The inventory class of a visual mark (kit, LED, board, overlay, etc.).
Factorization. Decomposition of an exposure into position, time, clarity, visibility, size, and viewership-related terms.
Pre-invest ROI. Decision support before capital and inventory are committed — productized further in Estimator.