Always-On incrementality

Incrementality you can act on, every day.

Stella turns every spend change into an automated causal study, then gives you daily incremental iROAS and response curves per campaign and channel. Insight to action in days, not the month a holdout takes.

  • Daily incremental iROAS per campaign and channel, with confidence bands.
  • Automated before-and-after causal studies on every qualifying spend change.
  • Most powerful when calibrated by your holdouts and MMMs, it inherits their accuracy.
Book a strategy callStart free

Included from Professional ($3,000/mo) and up.

Live demo, ask Stella about Always-On
Hi, I'm Stella.

Ask me about always-on incrementality: how incremental each channel is right now, where to move budget this week, and how Stella stays accurate between holdouts. Try one:

Trusted by growth teams who left enterprise measurement vendors

PopSockets
Momofuku
TUSHY
Rough Country
Neighbor
Pattern
Pirani
Plunge
WeightCare
Metro Vein Centers
Rylee + Cru
Miracle
PlantsBasically
Why always-on

Your incrementality is moving. A quarterly read cannot keep up.

Creative fatigues, audiences saturate, and competitors shift. The iROAS that was true in January is not true today. A holdout gives you a rigorous snapshot, but you make budget decisions every week. Always-On tracks how incremental each channel is, daily, so the number you act on is the one that is true right now.

Meta · daily incremental iROASupdated daily
3x2x1xWas 3.1x in January2.0x today
Campaign spend over time
Qualifying studyclean 15-day windowSkippedoverlapping change
How it works · Automated studies

Every spend change becomes a causal study.

When a campaign meaningfully changes spend, Stella runs a 15-day before-and-after causal analysis, controlling for your other spend and confounders. The change only qualifies if the window is clean, with no other change within a few days, so each read stays uncontaminated. A lite Bayesian model sets the baseline at the campaign level and rolls up to the channel.

How it works · Calibration

Automated reads count a little. Your holdouts anchor the truth.

An automated study on its own is directional, so Stella weights it low. Your calibrated MMMs count more, and your rigorous holdouts anchor the result. Always-On is most powerful when calibrated by those, because it inherits their accuracy between tests instead of guessing on its own.

How much each read counts toward the truth
Holdout experiment
anchor
Calibrated MMM
Automated daily study

Recent studies are weighted higher, so accuracy compounds as you keep testing.

Holdouts
Gold-standard geo experiments
MMM
Calibrated by your holdouts
Always-on
Live incremental factor, per channel
MTA
Weights every touchpoint
Ad platforms
Optimize for incremental margin
Holdouts
Gold-standard geo experiments
MMM
Calibrated by your holdouts
Always-on
Live incremental factor, per channel
MTA
Weights every touchpoint
Ad platforms
Optimize for incremental margin
↻ …and it loops, the more you run, the sharper the signal.
It gets smarter

The more you test, the more accurate it gets.

Recent studies are weighted higher, so accuracy compounds as you keep running holdouts and MMMs. Stella keeps two years of studies on hand and reweights continuously, so the daily number is always anchored to your most recent rigorous evidence.

Spend →Incremental returnHeadroomevery $ still worksSaturated$ past here is wasted
Daily response curves

Response curves per campaign and channel, every day.

See where each campaign and channel saturates, targetable to Shopify revenue, Amazon revenue, or total. You know which channels still have headroom to scale and which to pull back, this week, not next quarter.

What you can do

Daily, transparent, cross-platform.

Cross-platform incrementality

See how incremental Meta is on Shopify versus Amazon versus your total business, not a single blended number that hides the difference.

Campaign up to channel

Measure each campaign, then roll up hierarchically to the channel, so you can act at whichever level the decision lives.

Full transparency

Every automated study and its confidence band is visible, with two years of history. Nothing is hidden behind a single confident number.

What clients say

Hear it from the brands we run measurement for.

Jerel Blades
TUSHY

Always-On is included from Professional ($3,000/mo) and up.

How to work with Stella

Have us run it, or run it yourself.

Most brands want measurement done for them. That's our consultancy, and it's where most clients start. Not ready for done-for-you? The same platform is self-serve, from $6,000 down to free. There's an option for everyone.

The consultancy · from $10,000/mo

We build your causal optimization loop, and run it for you.

Proper holdouts and MMMs are real data science, and most teams aren't staffed for it. So a PhD data scientist builds your loop, runs it end to end, and hands you decisions and board-ready proof you can defend.

Book a strategy callDone right, without dedicating the headcount.
A model built and fit to your data
  • A PhD data scientist runs it end to end
  • Holdout & region/cluster design (even dense, multi-location)
  • Results validated until they’re decision-grade
  • Board / CEO / CFO-ready decks
  • Active strategy + a testing roadmap
  • Dedicated Slack + a named team

$10k isn't for everyone. There's a plan for every budget.

Prefer to run it yourself? Same platform, self-operated: the full loop at $6,000, proper holdouts from $3,000, or start completely free. Walk down to the plan that fits.

Free
$0/mo
For first holdouts & a feel for Stella.
  • Chat with your data: AI assistant for instant insights
  • 2 MMM runs / month
  • 2 incrementality runs / month
  • Manual CSV upload (no connectors)
Start free
Starter
$750/mo
Automated ingestion + unlimited experiments.
  • Automated state-level ingestion, up to 3 connectors (Shopify + Meta + Google)
  • Unlimited state-level geo holdout experiments
  • Unlimited regional (state-level) MMM
  • Unlimited manual data uploads
  • Chat with your data
Start free
Most popular
Professional
$3,000/mo
The full incrementality toolset.
  • Everything in Starter
  • Always-On Incrementality: daily iROAS, no holdout setup
  • Marginal growth curves per channel & campaign
  • Cross-platform halo effects (Amazon, Shopify, retail)
  • Automated DMA-level ingestion, up to 5 connectors + DMA-level holdouts
  • MCP: query your data from Claude Code & your AI tools
Start free
Scale
$6,000/mo
Everything, plus MTA & reverse ETL.
  • Everything in Professional
  • Multi-Touch Attribution (cross-channel journeys)
  • Reverse ETL: feed incremental conversions to Meta, Google & TikTok so they bid for incremental profit
  • Up to 12 connectors at DMA / zip-level volume
Start free
FAQ

Always-On incrementality questions, answered.

How does a spend change qualify for a study?

When a campaign meaningfully changes spend, Stella runs a 15-day before-and-after causal analysis, controlling for your other spend and confounders. The change only qualifies if the window is clean, with no other change within a few days on either side, so each read is not contaminated by something else moving at the same time.

Is automated measurement as accurate as a holdout?

On its own, an automated read is directional, which is why Stella weights it low. It becomes powerful when calibrated by your holdouts and MMMs: those rigorous results anchor the truth, and Always-On inherits their accuracy between tests. The more you test, the more accurate the daily numbers get.

Do I get results per campaign or per channel?

Both. Stella measures incrementality at the campaign level with a lite Bayesian model, then rolls campaigns up hierarchically to the channel level, so you can act on a single campaign or see the whole channel.

Which outcomes can I measure?

Target the response to Shopify revenue, Amazon revenue, or total revenue, so you can see how incremental Meta is on Shopify versus Amazon versus your blended business, not just one platform.

How fast is it compared to a holdout?

A geo holdout takes 20 days or more before you get a read. Always-On produces daily incremental iROAS and response curves continuously, so you move from insight to action in days. Run holdouts for the gold-standard number, and let Always-On keep it fresh in between.

What plan includes Always-On?

Always-On Incrementality is included from Professional ($3,000/mo) and up. Incrementality testing and media mix modeling are in every plan, free and up.

See your daily incrementality.

Book a strategy call and we will show you what Always-On surfaces on your channels, calibrated by the holdouts and MMMs you already run.

Book a strategy callor start free