By the Wavefield Research team · Published Sep 1, 2026 · Updated Sep 2, 2026
A political poll costs between roughly $2,000 and $40,000 or more, and the method sets the price. Live-caller statewide polls run $20,000 to $40,000+; IVR (robo) polls $5,000 to $10,000; an online district poll of 400 voters can field for about $2,500, most of it respondent costs.
Ask this question in public and you get two kinds of answers: forum threads quoting five-figure invoices from memory, and pollsters who would rather discuss it on a call. Both have a point — political polling cost really does span an order of magnitude — but the spread has causes you can price line by line. This guide does, using reported industry ranges and, for the self-serve path, our own published prices.
The honest unit is cost per completed interview, not the sticker price — a habit worth keeping from market research costs generally. Reported ranges: a high-quality national live-interview poll of 1,000 respondents runs $30,000–$40,000 (a figure pollsters have confirmed on the record), while online panel sample trades at single-digit dollars per complete. Between those poles sits every method choice a campaign makes.
Swipe the table sideways to see all columns →
| Method | Typical cost | Per complete | What you should know |
|---|---|---|---|
| Live caller (CATI) | $20,000–$40,000+ for a statewide poll; district polls commonly $15,000–$25,000 | $30–$100+ | The reference standard, priced by interviewer labor. Telephone response rates below 10% mean interviewers dial for hours per completed interview, and cell phones must be hand-dialed under US telemarketing law. |
| IVR (robo-poll) | $5,000–$10,000 per poll | a few dollars | Automated voice calls. Cheap and fast, but US law bars autodialed calls to cell phones, so IVR samples skew landline — older and less representative every cycle. Usually blended with an online supplement now. |
| Online with panel sample | $2,000–$3,500 for a 400-voter district study, all-in | $3–$8 general population; registered-voter targeting runs mid-to-high single digits | Panel respondents screened to your electorate, answering on their own devices. The economics that changed polling: sample is the main cost, and the platform fee is flat. |
| Online with lists you control | platform fee plus your channel costs | near zero for the sample itself | Party supporter files with email, membership lists, or a texting vendor with its own confirmed opt-in. The cheapest defensible poll there is — if the list honestly covers the electorate you claim to measure. |
Phone economics keep worsening for structural reasons: telephone survey response rates fell to 6% by 2018 in Pew Research Center's tracking and have not recovered. Every point of decline is more interviewer-hours per complete, which is why the live-caller column only moves in one direction.
Two polls of the same electorate can differ in price by 10x. These are the variables doing the work — and the first four apply whether a firm runs the poll or you do.
Cost scales linearly with completes; precision does not. 400 completes gives about ±4.9 points, 800 about ±3.5 — doubling the bill buys 1.4 points. District polls settle at 400–500 because past that, money usually serves subgroup reads better than the topline.
Every screened-out respondent still cost a contact. A likely-voter screen that passes half the sample roughly doubles the effective cost per usable interview, and a low-turnout primary electorate multiplies it further. Screens are the quiet budget line nobody quotes upfront.
Panels price by interview minutes and phone rooms by interviewer time, so a 20-minute instrument costs meaningfully more than a 10-minute one in any mode — and abandonment rises with length, which raises the contacts needed per complete.
A statewide sample is easy for panels to fill; a single state-house district is not. The smaller and more specific the electorate, the higher the per-complete price — and below a certain district size, voter-file outreach beats panels entirely.
Questionnaire design, fielding management, weighting, crosstabs, and the strategy memo are most of a pollster's invoice — the same anatomy as any agency quote. It is the part you can now unbundle: keep the counsel if you need it, automate the production.
The service layer deserves the closest look, because it is where a pollster quote and a platform bill part ways. On a $20,000 live-caller district poll, fielding labor is roughly half the money; most of the rest pays for the parts that feel like polling — questionnaire strategy, weighting to the electorate, crosstabs, the memo, the call where someone explains it. Those are real services. They are also, increasingly, software: questionnaire programming, raking to voter targets including past-vote recall, and significance-tested crosstabs are exactly the work an AI-driven platform automates. What software cannot replace is judgment about what to ask and what to do with the answer — which is the fair price of a good pollster, and the honest reason to hire one.
A concrete build: 400 likely voters in one district, a 12-minute instrument with a ballot test and a split-sample message battery, weighted to the district's demographics and past vote. Here is the whole political survey cost, both sample paths:
Platform: $199 (the Standard tier — weighting included, unlimited completes). Registered-voter panel sample at $5–$8 per quality-checked complete — arranged through Wavefield or bought direct from a vendor: $2,000–$3,200 for 400. Panel-vendor redirects with completion signing are built in, so quality control survives the handoff. Total: roughly $2,200–$3,400 — against $15,000–$25,000 for the same design commissioned live-caller.
Platform: $199, plus whatever your channels cost — a supporter file with email, a party membership list, or a texting vendor with its own confirmed opt-in (cold-texting a raw voter file is not consent, and our SMS invitations require documented opt-in). Total: a few hundred dollars, with one honest caveat: a supporter list measures your supporters. Claiming it measures the district is how internal polls end up wrong by ten points.
Either path produces the same deliverables: weighted toplines with margins of error on honest effective bases, significance-tested crosstabs, and exports. The full product story is on political polling for campaigns; the completes-to-precision arithmetic is in the sample size calculator.
In a contested race, the poll is an input to strategy — which message, which voters, what to concede. A good pollster sells judgment shaped by cycles of pattern-matching. If the readout will drive six-figure spending decisions, the counsel can be worth more than the fielding.
A poll released to press or donors borrows the credibility of whoever fielded it. A named firm with a public track record moves earned media in a way a self-run study cannot, however sound the method. Public-release polling is a reputation product; buy the reputation.
Some districts — very rural, very old, low broadband — are genuinely hard to reach through online panels. When the online frame misses your voters, live calls or a mixed-mode design stop being a luxury. Paying for reach beats saving money on a sample that cannot see the electorate.
Commissioned from a polling firm, a live-caller district poll typically runs $15,000–$25,000 and IVR $5,000–$10,000. Self-run online, the same 400-complete study costs the platform fee ($99–$299) plus respondents: $2,000–$3,200 for registered-voter panel sample, or almost nothing if you field to supporter lists you already control. For most local races, the poll is affordable; the discipline is the hard part.
400 completes is the working standard for a district poll (±4.9 points at 95% confidence), 600–800 for statewide (±3.5–4). Two caveats: the margin applies to the effective base after weighting, which is smaller than the raw count, and any subgroup you want to read separately — say, independents — needs enough completes of its own. Work it backward from the smallest group you must measure.
Not meaningfully. Political advertising rates collapse after Election Day because ad inventory is fixed supply, but survey interviewing is not: published panel pricing moves on incidence, interview length, and audience specificity, not the election calendar. Pollster availability improves in the off-season, which can help scheduling, but list prices for sample and platforms do not drop. If a poll is worth running, the calendar is a weak reason to wait.
Labor per completed interview. With telephone response rates in the single digits, an interviewer works through dozens of numbers for each complete, and US telemarketing law requires cell phones — most of the electorate — to be dialed by hand. Multiply an interviewer-hour by a 15-minute instrument and a likely-voter screen, and $50 per complete stops looking like margin and starts looking like arithmetic.
Related: political polling for campaigns · how much does market research cost · sample size calculator · survey screening questions
Brief the agent and it programs the ballot test, screens, and split-sample message battery; weighting and significance testing are built in. From $99 per study.
Start a study