Every cycle, thousands of down-ballot races are decided without a single poll — not because campaigns don't want data, but because a pollster engagement costs more than the media budget. Wavefield does the programming, fielding infrastructure, and statistics for $99–$299 a study. The races where a few hundred votes decide it can finally afford to know.
Free to build · $99–$299 to field · Tracker waves $99 · No retainer
A traditional polling engagement has three skilled-labor costs: a survey programmer to build the instrument, a project manager to babysit fieldwork, and an analyst to weight the data and cut the crosstabs. That is why it starts around $15,000 — and why targeting, message, and resource decisions in state-legislative races run on instinct and whatever the top of the ticket's polling implies.
All three of those costs are now automated. An AI agent programs the questionnaire, quotas manage themselves during fielding, and the weighting and significance testing a pollster's shop bills days for are computed the moment fielding closes. The question stops being which one race can afford a poll, and becomes which decisions you are still willing to make blind.
If the election for State Representative were held today, for whom would you vote?
Candidate order rotates per respondent; Undecided stays anchored last. Results come back weighted to district targets with the margin of error stated, not hidden.
The conventions of serious political research are defaults here, not add-ons — the same methodology covered in our weighting and sampling guides, enforced by the product:
Describe the study — likely-voter screen, ballot test, favorabilities, message battery, demographics — and the agent builds it in minutes: terminates, quota cells, candidate rotations, skip logic. You review every question before anything fields.
Rake results to district targets — age, gender, region, education, and past-vote recall, the dimension political samples skew on most. Weights are trimmed, effective bases computed, and every report shows the unweighted counts beside the weighted shares.
Randomized A/B arms are a first-class feature: half the sample sees message A, half sees B, assignment is uniform and automatic, and the crosstabs compare arms with significance testing. This is how message testing is supposed to work — not two separate surveys eyeballed side by side.
Every topline carries its margin of error computed on the Kish effective base after weighting — not the flattering raw count. A district study with n=300 has a real MoE and our reports state it. A poll that overstates its precision is a liability in a debate prep room.
Likely-voter screens terminate cleanly, quota cells fill and close automatically, and respondents cannot tell a screen-out from a completion. Interrupted respondents can resume where they left off, attention-check items flag the inattentive, and field monitoring shows completes, incidence, and quota progress live.
SPSS files with weights registered, crosstab banner books in Excel, and the topline as an editable Word document. Tracker waves re-field the identical instrument for $99 and the trend report compares waves with significance tests.
Political data lives or dies on credibility, so here is exactly where the platform's job ends and yours begins:
Wavefield fields through shareable links, single-use email invitations, and consent-based SMS. Campaigns reach voters through their own channels — supporter lists, volunteer networks, existing outreach — or pair the platform with panel sample you source; panel-vendor redirects with completion signing are built in. We do not sell sample, and we will not pretend otherwise.
Our SMS invitations require confirmed opt-in — that is the compliance story, and it means a raw voter file with phone numbers does not qualify. Campaigns that text voter files use dedicated political texting vendors under their own compliance regimes; studies here field through channels where consent exists.
A state-house district study is often n=250–400. That is workable — campaigns decide with it every cycle — but the margin of error is real and ours is printed on the report. If a result is inside the margin, the report will not let you pretend it is a lead.
A full-service pollster brings judgment: what to ask, when to field, how to read a shift. The platform does the programming, fielding infrastructure, and statistics at a price that makes data possible where it was not — the strategic reading stays with your team or your consultant.
Delivery channels and their trade-offs — email, SMS, links, panel — are covered in survey data collection methods; the SMS consent mechanics are documented on the SMS surveys page.
That is the point of per-project pricing. A traditional pollster engagement commonly starts around $15,000–$30,000, which excludes most down-ballot races. A Wavefield study is $99–$299 depending on complexity, with repeat tracker waves at $99 — so a committee overseeing twenty target districts can afford data in every one of them for less than one traditional engagement.
From you. Campaigns field through shareable links, single-use email invitations to lists with consent, and SMS invitations where opt-in exists — or through panel sample the campaign sources from a panel company, using the built-in vendor redirect support. Wavefield does not supply respondents, and any political poll's credibility starts with being honest about where its sample came from.
Yes. Weighting targets are yours to define, and past-vote recall is a standard dimension alongside age, gender, region, and education. The raking engine handles the iteration, trims extreme weights, and computes significance on effective bases — the same conventions a professional polling shop applies.
The questionnaire is programmed in minutes from a brief, and review is the only step between you and fielding — same-day launch is normal. Fieldwork pace depends on your channels: an engaged supporter list can return hundreds of completes overnight, and the weighted topline is ready the moment fielding closes.
Work out the completes a district study needs with the sample size calculator · Full analysis workflow: how to analyze survey results
Brief the agent, review the questionnaire, launch to your channels. Free to build — you pay only when you field.
Start a study