Strategy with receipts. Confidence-scored marketing decisions, with the work included.
Voltaire's executive tier. Ranked hypotheses, explicit tradeoffs, decision log, monthly memo. The strategy a senior marketer would deliver if a senior marketer could read every relevant case study every week.
Most agencies sell you a plan. Voltaire sells you a continuously updating decision.
Execution speed is no longer the bottleneck. Allocation is. With finite budget and a market that changes every month, the harder question is not how fast you ship. It is which positioning, which channel mix, which experiment, which brand bet to run in the first place. Decide answers that question, with confidence intervals, every month, and tracks its own accuracy openly.
Four deliverables. Every month.
Strategy Memo
Monthly. 5 to 15 pages. Three to seven ranked recommendations with confidence intervals, evidence trail, and a tradeoff matrix scored against multiple objectives (3-month revenue lift, 12-month brand asset accretion, payback period, downside risk). You pick which objective matters more.
Decision log
Live web. Every recommendation timestamped. Every outcome recorded against it as the work executes (by Voltaire or by your team). Retrospective accuracy tracked openly. We hold ourselves to the receipts.
Experiment spec
Monthly. The cheapest test that meaningfully discriminates between your top two hypotheses, with predicted information gain per dollar. Real campaigns become evidence-gathering operations, not bets.
Quarterly executive briefing
60-minute recorded session and a 5-page written summary. How the system is learning your business, what it now believes with high confidence, and what next quarter's portfolio looks like.
Inputs, process, outputs. No magic.
Inputs
- Your first-party data via authorized integrations: Stripe revenue, GA4 traffic, your Meta and Google ad accounts, your CRM (HubSpot, Pipedrive, Attio), your email tool (Resend, Loops, Klaviyo).
- Voltaire's existing AI search and SEO signal: AEO citation tracking across ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews. Competitor content and ad teardown.
- Public competitive data: Meta Ad Library, Google Ads Transparency, SimilarWeb, LinkedIn company insights, Crunchbase profile.
- Prior knowledge from a corpus of public marketing case studies, agency reports, MMM literature, and podcast transcripts, indexed for retrieval.
Process (every month)
- Generate three to seven candidate strategies for your business based on the signal we collected and the priors we cite.
- Score each by confidence, with explicit ranges grounded in the evidence behind them. Every claim traces back to a source.
- Expose tradeoffs against the multiple objectives you care about, with the weights you set. We do not pretend marketing has one objective function.
- Design the cheapest experiment that would tighten our confidence on your top two hypotheses, ranked by information gain per dollar.
Outputs
- The Strategy Memo, the decision log, the experiment spec, the quarterly briefing.
- Recommendations the agency execution loop (or your team) can ship next week, not next quarter.
Honesty about what is in the box, and what is on the roadmap.
A federated network of comparable businesses sharing outcome data.
We are building this with the customers we onboard. We are not selling it as if it exists today.
A consumer simulation engine.
On the roadmap. Useful for cheap pre-tests of campaign hypotheses before real spend. Not in the Decide box yet.
A formal causal transfer model.
Phase 2 of our internal roadmap. Today's Decide Memos are powered by retrieval over priors plus your first-party data. The model layer comes once the network has the volume to train it.
A claim of optimal marketing.
Marketing has no single objective function. We rank hypotheses under the objectives you set. You pick the weight. Anyone selling you the optimal strategy is selling you a number that does not exist.
Sales-led. No self-serve checkout.
Confidence-scored strategy at this price-point requires a buyer conversation about objectives, decision rights, and integration scope. We start with a 30-minute call.
For buyers with existing execution who want the decisioning layer.
- +Monthly Strategy Memo (5 to 15 pages, ranked recommendations, confidence intervals, tradeoff matrix)
- +Live decision log with retrospective accuracy tracking
- +Monthly experiment spec ranked by information gain per dollar
- +Quarterly executive briefing (recorded + written summary)
- +Senior strategist supervision on every memo
- +First-party connector pack across your stack
Decide plus the full execution loop. The default for new buyers who want the whole system. Saves $397/mo vs buying separately.
- +Everything in Decide, plus:
- +Discover module (AI search, SEO, competitor signal)
- +Capture module (content production, paid acquisition, email, landing pages)
- +Compound module (voice calibration, brand identity, attribution loop)
- +Single accountable team across strategy and execution
Multi-region, multi-brand, dedicated strategist plus account executive, real-time dashboard, SLA-backed delivery.
- +Everything in Decide + Loop, plus:
- +Multi-region citation and content support
- +Multi-brand portfolio coordination
- +Real-time dashboard and API access
- +SLA guarantees on memo delivery and experiment turnaround
- +Dedicated strategy partner
Every recommendation is timestamped. Every outcome is recorded against it.
The decision log is the proof. We track our own accuracy openly. Monthly billing is available, so you can step out at any cycle. The longer commitments are there if you want the price discount, not because we want to lock you in to a strategy that is not working.
Decide, in detail.
Discover collects signal: where buyers are looking, what competitors are doing, where citations are leaking. Decide synthesizes signal into ranked bets with confidence intervals. Discover tells you what is happening. Decide tells you what to do about it, and how confident you should be in each option. Decide consumes Discover's output as one of its inputs.
The self-serve cart sells autonomous execution at software margins. Decide adds senior strategist supervision, a confidence-scoring framework, and a decision log Voltaire is publicly accountable to. The price reflects the human supervision and the accountability. Buyers who only need execution stay on the cart. Decide is for buyers who need allocation answers under uncertainty.
Every recommendation is timestamped in your decision log. Every outcome is recorded against it as the work executes. We expose our accuracy openly in the log itself, and the quarterly executive briefing covers what we got right and what we got wrong. Monthly billing is available, so you can step out at any cycle. The longer commitments are there for the price discount, not for vendor lock-in.
No. Decide can hand recommendations directly to your existing in-house team or current agency. The Decide + Loop bundle saves $397 per month if you want both strategy and execution under one roof, and it removes a coordination layer.
Stripe (revenue), GA4 (traffic and conversion), your Meta and Google ad accounts, your CRM (HubSpot, Pipedrive, or Attio), and your email tool (Resend, Loops, or Klaviyo). Read-only access via standard OAuth. Your data never leaves your tools. Voltaire reads through authorized integrations and does not scrape or store PII.
Connector pack live in week 1. First Strategy Memo by end of week 4. Decision log live from day one. By month 3 the loop has accumulated enough data on your specific business that the memos sharpen visibly compared to month 1.
Real, and intentionally wide where the evidence is thin. We do not pretend to know things we do not know. A typical recommendation might read: we estimate this campaign lifts pipeline 8 to 22 percent with 80 percent confidence, and the cheapest experiment that would tighten this to 12 to 15 percent is described in the experiment spec.
No. Decide is the executive tier of the same AI-native marketing agency. Same team, same accountability, same brand. Decide is the strategy stage of Voltaire's demand engineering loop. The other stages (Discover, Capture, Compound) remain available as the self-serve cart.
Ready to see what a confidence-scored strategy looks like for your business?
30 minutes with Fabrice. No pitch. We map your inputs, your objectives, and your decision rights, then tell you whether Decide is the right fit.