Cost per click · Google Search
Car insurance quotes · today
Predict the market.
Move
first.
Forecast
Budget moved · Meta to Google Search
RatioCast campaign · executed 14:15
- Change
- +$2,400 today on Google Search, −$2,400 on Meta
- Why
- Next-dollar return forecast at 1.7× on Google, 1.1× on Meta. A rival’s budget is pacing to run out by 15:30.
- Expected
- +3% to +7% conversions today, 80% confidence
- Limits
- Within max budget $40,000 a day and min tROAS 350%
Holdout test · conversions per $1,000
RatioCast campaign against its holdout · day 9 of 28
Algorithmic trading for paid media. RatioCast reads your ad accounts, first-party data and market signals in real time, and moves budgets, targets and bids toward your real goal, before the market does.
Markets move in minutes. Marketing, in weeks.
Ad auctions reprice by the minute. Yet every advertiser now runs the same AI agents on the same clock.
Conversions report back in days, and most teams edit campaigns, budgets and targets weekly at best.
86% of marketing teams already use AI. Meta’s Advantage+ runs at a $75B+ run-rate, and over 30% of Google Search spend is in AI campaigns.
Same data, same AI, same clock. The edge is knowing where the auction goes next.
How fast things move
Each figure links to its source
Marketing, even with AI
Ad auctions are markets
Trading went algorithmic decades ago. RatioCast brings that discipline to marketing, and treats every auction like the market it is.
- Exchange
- Google, Meta and TikTok auctions
- Price
- CPC and CPM
- Volatility
- Swings in CTR, CVR and CPC
- Order-book depth
- Bid headroom and impression share
- Other traders
- Competitors entering and leaving
- Macro events
- Seasonality, news and weather
- Alpha decay
- Creative fatigue
- Broker as counterparty
- The platform sells the inventory and bids for you
- Backtesting
- Simulate first, then run holdouts
One loop, in real time
Don’t act on what happened. Forecast what’s next. Swipe or use the arrows to follow the loop.
Data feeds
Streaming
| Source | Updated | Rows today |
|---|---|---|
| Google Ads | 2 min ago | 48,210 |
| Meta | 4 min ago | 31,906 |
| TikTok | 6 min ago | 12,447 |
| Site conversions | 1 min ago | 3,118 |
| Weather, news, trends | 11 min ago | 624 |
Auction X-ray
Google Search · car insurance quotes
| Rival | Imp. share | Budget left | Signal |
|---|---|---|---|
| Rival A | 31% | 26% | Runs out ~15:30 |
| Rival B | 18% | 71% | New since 09:40 |
| Rival C | 14% | 58% | Steady |
| Others | 22% | n/a | No change |
Event forecast
Demand index against a normal day
Known event: Black Friday. Forecast the surge, move first.
Unknown event: wildfires lift insurance demand. Spike caught within the hour.
Return on the next dollar
Forecast for the next 6 hours
Move $2,400 from Meta to Google Search until 18:00.
Your limits
Applied to every RatioCast action
| Limit | Set to | Today |
|---|---|---|
| Max budget | $40,000 / day | $31,850 |
| Min tROAS | 350% | 412% |
| Max tCPA | $85 | $71 |
| Holdout share | 10% | 10% |
Sense
Your ad accounts, your first-party data and external market signals, read in real time and corrected for conversions not yet reported.
Real-time data pipeline
Your goal, not the platform’s
Platform bidders chase a ROAS or CPA target on average. RatioCast steers them toward what you actually want, and every move it makes is measured against that goal.
Pick a goal to see how RatioCast steers.
- Maximises
- Profit: revenue times margin, minus ad spend
- Steers
- Sets each campaign’s tROAS where the next dollar still earns more than it costs
- Guardrail
- Never below your min tROAS of 350%
Proposed Google Search tROAS 410% → 370%. Forecast: the next dollar still returns 1.4× after margin.
- Maximises
- New customers, at or above your return floor
- Steers
- Raises budgets and loosens targets wherever the next dollar clears the floor
- Guardrail
- Return floor 300%, max budget $40,000 a day
Proposed TikTok budget +18% today. Forecast: return stays at 330%, above the floor.
- Maximises
- Users who convert and stay, at a lower cost each
- Steers
- Feeds your first-party quality signals back, so bidders favour users like your best customers
- Guardrail
- Max tCPA $85
Proposed Meta tCPA $85 → $72, weighted to high-retention audiences.
Illustrative example. Your goal and limits are set together at onboarding.
From the calendar to the forecast
DecisionToday, on the calendarWith RatioCast, on the forecast
GoalThe platform’s target: ROAS or CPA on averageYour goal: profit, growth, or cheaper and better users
Channel splitYearly, from the media mix modelDaily, toward the best next dollar
tROAS and tCPAQuarterly, on the averageDaily, on the margin
CreativeChanged after results dropChanged before fatigue hits
CompetitorsSeen in next week’s reportForecast by the hour
Unknown eventsMissedWithin the hour: a rival pulls out, a trend breaks, a checkout fails
ProofPlatform-reported resultsMeasured against holdouts
One campaign first, then the whole budget
RatioCast earns the budget step by step. It never spends outside the limits you set, and every result is read against a holdout.
-
First, connect your accounts. RatioCast reads them in real time, with your first-party data and external market signals.
Your ad accounts
- Google Ads
- Meta
- TikTok
First-party data
- Site conversions
Market signals
- Trends
- Weather
- News
Reading your accounts, first-party data and market signals in real time.
-
Then, set the limits. RatioCast runs one campaign per channel inside them.
RatioCast campaigns created on Google Search, Meta and TikTok. Capped by these limits.
-
Finally, prove it against a holdout, then expand to the full budget.
Land
One capped campaign per channel
DoneProve
Lift measured against a holdout
Day 9 of 28Expand
The full budget, once it is proven
Next
Under the hood
Four capabilities that are rare on their own. RatioCast runs all four together, for every account.
A note from the founder
Tzahi Zilbershtein
Founder & CEO
I have spent my career building the engines that bid in ad auctions. I led an engineering group in Google Ads EMEA, served as CTO and VP R&D, hold patents on auction bidding systems, and hold an Executive MBA.
RatioCast starts from one observation. AI now runs the bid for everyone, on the same data and the same clock, but it does not look ahead. Budgets, targets and the channel split still move on the calendar. Trading desks solved this decades ago with models that forecast the market and act inside strict risk limits. RatioCast brings that discipline to ad auctions: forecast, act inside limits, measure against a holdout.
The AI behind RatioCast is of a specific kind. Not a chatbot, but sequence and time-series models that learn how each auction moves, reverse-engineer rivals from partial signals, and get sharper with every account they learn from.
We are starting with a small number of large performance advertisers on Google, Meta and TikTok. If that is you, I would like to show you what RatioCast sees in your auctions.
See the road ahead
Smart Bidding is the engine. RatioCast is the driver who sees the road ahead.