The Supply-Side Alt-Data Gap: Why Yipit and Second Measure Stop at the Consumer

Alternative data is a $5bn+ annual market. The vast majority of that spend covers the consumer side: what people buy, where they go, how much they spend. Yipit Data, Second Measure, Bloomberg Second Measure, Placer.ai, App Annie, billion-dollar businesses built on the premise that consumer behavioral signals predict corporate earnings before the earnings report.

The supply side has nothing equivalent. This post explains why, what that gap is worth, and why AI voice agents are the first technology capable of closing it.

What Yipit and Second Measure actually do

Consumer-side alt-data companies ingest anonymized behavioral signals, credit card transactions, loyalty program data, geolocation pings, app telemetry, and aggregate them into structured panels. The output is a recurring dataset: "Walmart same-store sales up 3.2% week-over-week, based on 2.4 million transactions."

This works because consumer behavior is:

• High-frequency. Purchases happen daily. The signal is continuous.
• Passively collectable. The data exhaust of consumer activity (transaction logs, location pings) exists; the alt-data company just aggregates and licenses it.
• Directly tied to reported metrics. Consumer transaction data predicts same-store sales, GMV, active-user counts. The feedback loop from data to prediction to outcome is tight.

The supply side doesn't have these properties. What a brand partner thinks about Zalando's payment terms, what an H&M factory supplier thinks about order visibility, what a restaurant franchisee thinks about corporate's pricing strategy, these signals don't exist as passive data exhaust. They exist only in people's heads, and they come out only when someone asks.

The supply-side signals that matter

For a European fashion e-commerce investor, the signals that predict platform outcomes exist at the supply side, not the consumer side:

For Zalando specifically:

• Are top-tier brand partners planning to reduce allocation in favour of their own DTC channel or a competitor marketplace?
• Are payment terms getting worse, signaling Zalando is using its scale to extract working capital from brand partners?
• Are brands reporting that marketing support ROI has declined, suggesting rising CAC or declining platform traffic quality?
• What percentage of brands say they'd recommend Zalando to a peer brand? That NPS compares to ASOS, About You, Amazon Fashion.

None of these appear in Zalando's own reporting. The closest proxy, brand partner GMV growth, is reported with a 60–90 day lag, aggregated, and subject to platform framing. The supply-side signal is primary, forward-looking, and independent.

The same logic applies across verticals:

• H&M's garment factory suppliers know whether lead times are deteriorating, whether H&M is paying on time, and whether they'd prefer to be sourcing to a competitor retailer. This information predicts supply-chain disruptions before they hit public disclosures.
• Restaurant franchisees know whether corporate's new pricing strategy is generating pushback, whether staff retention is getting harder, and whether comp-sales feel like they're declining. This predicts what the franchisor will report in 90 days.

Why the gap persisted for so long

Three structural barriers prevented supply-side panels from reaching scale:

1. Economics. At €50–200 per interview (GLG, AlphaSense, Tegus), a 5,000-respondent quarterly panel costs €250,000–€1,000,000 per wave. That's not a dataset; it's a bespoke consulting project. No alt-data business can build a recurring, institutionally-subscribed product at those unit economics.

2. Language and geography. European supply-side research requires German, French, Dutch, Turkish, Bangla, and Hindi. Human-caller operations at this multilingual scale require expensive staffing in each market.

3. Consent and disclosure. AI-voice-agent calling requires upfront AI disclosure (per GDPR Art. 22 and emerging global standards). Until 2024, the voice quality of AI agents was insufficient to sustain a real-time structured interview without high abandonment rates.

AI voice agents resolved all three. Cost is now €1–2 per completed interview at scale. Multilingual agents run natively. Voice quality improved sufficiently to sustain 4–6 minute structured interviews with low abandonment.

The three-sided market

The supply-side alt-data market is not just hedge funds. It's structurally three-sided:

Side 1: Measured platforms (defensive buyers)

Zalando cannot run a brand-partner satisfaction survey themselves. The commercial relationship prevents honest answers. Independent research fills the gap, and the platform pays to know what its own partners really think. Analogues exist in every adjacent sector: Glassdoor for the employer-employee relationship, J.D. Power for the dealer-OEM relationship.

Side 2: Competitor platforms (offensive buyers)

ASOS wants to know where Zalando is weak, which brand categories are most dissatisfied, which geographies have the worst payment terms, which brands are most likely to be attracted by a better offer. This is standard competitive intelligence. The insight value is equivalent to an expert network engagement at 1/100th the cost.

Side 3: Brand houses (allocation decision buyers)

Adidas has €Xbn in wholesale allocation to distribute across Zalando, ASOS, About You, Amazon Fashion, and its own DTC channel each season. The brand-partner sentiment data tells Adidas where it will get the best payment terms, the best marketing support, and the highest likely sell-through. This is a CFO-adjacent data purchase, not just an analyst tool.

What the addressable market looks like

One dataset per vertical. Three buyer sides per dataset. Enterprise contracts in the €50,000–€300,000/year range per buyer.

For European fashion e-commerce alone (BPI):

• ~10 platform-side buyers at €100,000–€300,000/year each
• ~50 brand-house buyers at €50,000–€100,000/year each
• ~20 hedge fund / alt-data buyers at €30,000–€80,000/year each

That's a €5–15M ARR addressable market for a single vertical dataset with three buyer sides. The catalog extends to any B2B or B2C market where one side of the market has unexpressed opinions about the other.

Brand Partner Index

Brand Partner Index is Auraqu's entry point into supply-side alt-data. Quarterly, independent, AI-voice-agent-collected sentiment from ~5,270 European fashion brand partners. The first systematic dataset of its kind.

Vol. I preview targeting Q2 2026. View the publication or contact us for institutional access.