When the Forecast Meets the Quarter

A half-year of digital risk, read as a sequence: what the first quarter predicted, and how the second answered.

ICDS DIGITAL RISK REPORT

OPEN RESEARCH SERIES

PUBLICATION

Digital Risk Report · H1 2026

PERIOD COVERED

First half 2026 · Q4 2025 baseline · Q1 forecast · Q2 tes

SERIES

Open Research Series · Institute of Communication and Data Science

DATA VOLUME

87,580 reactive-layer mentions across four sectors · Q2 2026

SECTORS TRACKED

Healthcare · Finance / Fintech Professional Services · Retail (control)

Vadim Matyushkin

Institute of Communication and Data Science

Director and Lead Researcher

PREFACE

Every January, the risk industry receives its seasonal barometer: a survey of hundreds of experts on what they expect from the world over the next two and ten years. It is a valuable instrument — it shows where the attention of decision-making elites currently points. But it answers one specific question — what people believe about risk — and leaves a different question unanswered: what risk is already doing while that belief is still forming.

We do not ask experts what they think. We read the trace those risks are already leaving in public communication — before it becomes an official damage figure, a lawsuit, or a line in a quarterly filing. A forecast made in one quarter must be either confirmed or overturned by the next quarter's data — otherwise it isn't a measurement.

"We occupy the space where risk hasn't yet become an opinion, but has already become a signal."

Key Findings · H1 2026

Four sectors. Three forecasts confirmed. One that could not be forecast — and that is itself the finding.

Q1 2026

Was the half-year's turning quarter

In finance and professional services, the reactive layer surged and the masking effect peaked — the widest gap yet recorded between how a sector sounds and how it actually reacts. Q1 did not just describe risk; it produced forecasts that Q2 either confirmed or overturned.

71%

Healthcare reactive layer Q1 — forecast held

A shift from calm to event-driven flagged in Q1 was confirmed by a run of overlapping incident waves in Q2. The reading held — and also exposed the method's current limit: overlapping events resist clean separation.

−88

Finance reactive tone beneath a −35 surface

Q1 masking reached 60%: observed tone read −35 while the reactive layer ran −88, roughly 2.5× sharper. Q2 confirmed the sharpness — reactive negativity reached 45%, the highest of the half-year.

−114

Professional Services — widest mask, confirmed

Q1 observed tone −42 against reactive tone −114. Q2's law-firm attack series bore the reading out almost exactly — negativity held steady at 42%.

57%

Retail — from silence to first signal

Silent through Q4 and Q1 with zero reactive layer, retail produced breach-driven incidents in Q2 — 57% negative, highest share of any sector. Silence cannot be forecast; events can.

About This Issue

ICDS Digital Risk Report · H1 2026 · Open Research Series

Q4 2025 and Q1 2026 figures are drawn from the full two-layer corpus. Q2 2026 figures are drawn from the reactive layer alone. Absolute volumes are not compared across that boundary — negativity share, tone, and event structure are.

3 periods

Q4 2025 baseline · Q1 2026 forecast · Q2 2026 test

4 sectors

Healthcare · Finance / Fintech · Professional Services · Retail (control)

2 layers

Inertial and Reactive — measured and reported separately

87,580

Q2 reactive-layer mentions across four sectors

What’s Inside

The report follows a single arc: baseline, forecast, examination. Each section takes one sector or one question through all three.

Preface

Why this report is built as a sequence rather than a sum — and what it means for a method to be falsifiable within its own cycle. By Vadim Matyushkin, Director and Lead Researcher, ICDS.

02

Overview of Methodology & Key Findings

How the two-layer corpus works, what masking measures, and why Q4/Q1 and Q2 figures are not compared on volume.

The Half-Year as a Sequence

Q4 gives resting temperature, Q1 produces forecasts, Q2 is the examination. A single quarter is a photograph; three in a row is a trajectory.

04

The Trajectory, Sector by Sector

Comparative table across all four sectors: Q4, Q1, and Q2 negativity share plus trajectory reading.

Sector Profiles: Healthcare · Finance · Professional Services · Retail

Each sector taken through the same arc — rest, forecast, and what Q2 returned. Includes named method limits where they applied.

06

Forecast and Verification

The report's spine: Q1 readings mapped against Q2 verdicts. Three confirmed. One changed in a way no forecast could anticipate.

What Is Claimed, and What Comes Next

What the half-year establishes and what it does not. Named limits. The open questions Q3 will answer — or fail to.

08

Appendices & Methodology Note

Appendix A: Education — a fifth sector in view. Appendix B: glossary. Methodology note on sources, comparability, and data boundaries.

"A total for the half-year would show retail at 57% negative and stop there; the sequence shows it silent for two quarters and then, suddenly, not — which is the part worth acting on."

Digital Risk Report · H1 2026 · Section 7: Forecast and Verification

Forecast & Verification

The report's spine — a sequence, not a sum

Overlapping incident waves through late April; weeks of 71% and 67% negative. The reading held — and exposed the method's limit: events arriving too close together resist clean separation.

Sharp reaction masked by volume. Observed tone −35, reactive tone −88 — 2.5× sharper than the surface showed. Masking 60%.

Sharp reaction masked by neutrality. Negativity nearly doubled from 24% to 43%. Reactive tone −114 against observed −42 — widest mask of the four.

No forecast possible. Silent control — zero reactive layer through Q4 and Q1. Nothing to predict in the absence of an event.
noise but by breach events with names attached.

Turning event-driven. Reactive layer 71% of volume — most of what the sector carried was no longer background but reaction to events.

Professional Services

Retail

Finance / Fintech

Healthcare

Sector

CONFIRMED

CONFIRMED

CONFIRMED

CHANGED

Reactive negativity reached 45%, highest of the half-year. Public companies disclosed incidents while playing them down — opening the gap the method is built to catch.

Acquired a breach-driven signal of its own. Week of 18 May: 90% negative. Across Q2: 57% negative — highest share of any sector, driven not by ambient noise but by breach events with names attached.

Law-firm attack series held the reading almost exactly at 42%. Caution: weakest reach coverage of any sector; readings rest on a thinner base.

Q1 Reading

Q2 Verdict

Q1 2026 · BRI ACROSS SECTORS · PREVIEW

The Trajectory in Numbers

Healthcare

10%

38%

45%

42%

57%

25%

Rising each quarter · forecast confirmed

Finance / fintech

24%

35%

Masking 60% · sharp reality beneath calm surface

Professional services

24%

43%

Widest mask · Q1 reading held

Retail

48%

43%

Q4 2025 and Q1 2026 from the full two-layer corpus. Q2 2026 from the reactive layer only. Negativity share is comparable across all three quarters; absolute volumes are not.

Silent control → first breach signal in Q2

Who Should Read This Report

The H1 report is addressed to four professional audiences. Each finds a different signal in the same sequence.

RISK MANAGEMENT

Risk Managers & CISOs

The H1 sequence shows which sectors moved from stable background to event-driven — and how early the signal appeared before incidents became official. A reading made in one quarter was testable in the next. Healthcare's reactive layer hit 71% in Q1; the incidents that justified it arrived across Q2.

Healthcare reactive layer Q1: 71% · Q2 confirmed · Professional Services mask 63% · reading held at 42%.

INSURANCE & RISK PRICING

Insurers & Actuaries

The reactive layer surge is a leading indicator of claims with a 1–2 quarter lag. The H1 sequence provides the first empirical test of that claim: Q1 masking figures and reactive-layer surges anticipated what Q2's incidents confirmed in three sectors. The sequence is the evidence — not any single quarter.

Finance reactive tone Q1: −88 · Q2 reactive negativity: 45%, highest of H1 · Retail: from zero reactive layer to 57% negative.

INVESTMENT & IR

Investment Analysts & IR Teams

Finance carries the half-year's most sensitive forward indicator: protective behaviour — the moment an audience stops discussing a threat and starts looking for ways to guard against it. Q2 added one observation: public companies disclose incidents while playing them down, opening a gap between what a company admits and what the external signal shows.

Finance observed tone: −35 · reactive tone: −88 · gap confirmed in Q2 · protective behaviour signal: direction under active test.

COMMUNICATIONS

Corporate Communications Leads

The H1 report documents the gap between what standard monitoring shows and what is actually moving audiences — and then verifies that gap against the following quarter. Masking is not an obstacle to measurement here; it is the object of measurement.

Professional Services: observed −42 · reactive −114 · masking 63% — widest of the four sectors across both quarters.

Request the Full Report

The complete H1 2026 Digital Risk Report includes the full sector arcs, forecast verification table, methodology notes, and appendices — including a fifth sector in view for the next issue.

14 pages · 4 sector profiles · forecast & verification table · open research series

Digital Risk Report · H1 2026 · First half 2026
Prepared by Institute of Communication and Data Science (ICDS) · Open Research Series · icds.institute