The Brief. In April 2026, organizers including the People’s Union USA and Black Voters Matter (“We Ain’t Buying It”) named Amazon, Target, Walmart, Home Depot, Starbucks and CBS as boycott targets. The question every brand on that list is asking: how many of these threats are real?
Surveys can’t answer it. Ask “would you boycott Target?” and people perform a value — they don’t simulate a Tuesday. So we ran the same boycott question two ways across a 570-twin US panel: once as a stated-intent (SAY) question, and once as a simulated-behaviour (DO) prompt that drops the twin into a normal week with real friction — the cart already full, the nearest alternative 20 minutes further. The within-twin SAY−DO delta is the say-do gap.
| Panel | 570 US digital twins, stratified by political lean, generation and news diet |
| Brands tested | Target, Amazon, Walmart, Home Depot, Starbucks, CBS — each with the real 2026 trigger as scenario stimulus |
| Design | Say-Do Gap: same question in SAY mode (stated intent) and DO mode (simulated normal week). Neutral vs activist framing as paired arms. |
| Outcome metric | Stop or meaningfully cut spending (yes / no / don’t know) |
| Run date | June 2026 |
Every survey playbook warns that stated intent over-states action. Here, the opposite happened — and it’s more dangerous for brands.
When asked to state an intention, three out of four twins took cover in “don’t know.” Boycotting is a values question, and the safest answer to a values question is to not commit. But the moment the same twins were dropped into a real week — cart full, trigger known — the fog lifted: “don’t know” fell from 73.5% to 11.8%, and the real boycott rate came in above the stated one, at 22.0%.
Of every twin who said “don’t know,” 75.8% quietly carried on shopping, 11.1% actually boycotted, and 13.1% stayed undecided. Meanwhile 79.6% of the people who said they’d boycott followed through. Stated intent isn’t noise — but the “don’t knows” are, and a survey that reads them as soft boycott support is reading them backwards.
The lesson isn’t “people lie.” It’s that the stated-intent question has no answer for most people until a real situation forces one. The DO simulation forces it.
Simulated (DO) boycott rate per brand. This is the number that maps to real-world foot traffic — not the survey number.
Target sits at the top at 32.9% — and that isn’t a panel artefact. Target’s real-world foot traffic fell roughly 16% after it rolled back DEI and its $2B Black-owned-business pledge. The panel is reading the same signal the parking lots did. At the other end, Amazon and Walmart barely move: their boycott threats are loud in the discourse and nearly inert in behaviour.
Brands get boycotted for betrayal, not just behaviour. Target and Starbucks made values-promises and were seen to break them — that reversal is the trigger. Amazon and Walmart never positioned on those values, so there’s no promise to betray, and no substitute convenient enough to make leaving worth it.
DO boycott rate by the kind of news a twin consumes (joined on person ID from a separate media-diet study; 431 of 570 twins matched).
The single sharpest divide isn’t left vs. right — it’s engaged vs. ambient. Twins on a left / legacy-news diet boycott at 44.3%, more than double the 19.5% of the social-only majority. The 138 social-only twins and 107 no-news twins — together the largest share of the panel — are the brands’ firewall: they may hear the outrage, but it doesn’t convert to a changed cart.
A boycott’s real reach depends less on how angry the loudest segment is and more on how much of the customer base lives outside the news cycle entirely. For Target, the engaged-news segment is large enough to dent traffic. For Amazon, even the 44% Left/legacy boycott rate (24% for Amazon specifically) drowns in a customer base that mostly doesn’t track the story.
| DO boycott % by brand × news diet | Right | Left/legacy | Social only | No news |
|---|---|---|---|---|
| Target | 26% | 62% | 28% | 30% |
| Starbucks | 28% | 48% | 28% | 30% |
| CBS | 35% | 54% | 25% | 24% |
| Home Depot | 17% | 41% | 16% | 21% |
| Walmart | 7% | 36% | 11% | 15% |
| Amazon | 15% | 24% | 9% | 17% |
Same brand, same twin pool, same DO simulation — only the framing of the trigger changed (neutral fact vs. activist call-to-action).
Framing isn’t just an opinion-mover — it changed simulated behaviour. Presenting the trigger as an activist call-to-action rather than a neutral fact lifted the DO boycott rate by 9.4 points. And it isn’t a one-tribe effect: the lift held at +10.7 for Right-news twins, +10.0 for Left/legacy, +9.6 for no-news, +7.4 for social-only. The organizer’s framing is doing measurable work.
For a brand, this is the volatility term: the same underlying grievance can land at 17% or 27% depending entirely on whether an organized campaign frames it. Brands on the list should forecast the activist number, not the neutral one, when a coordinated campaign is live.
Public benchmarks line up with the panel: YouGov found 48% say they’d boycott vs. 38% who actually did; Target foot traffic fell ~16% after the DEI rollback; and participation typically fades after ~8 months. The say-do gap, the Target signal, and the decay are all visible in the synthetic panel — before a single real customer was surveyed.
