For teams that need to know what worked

Did your change really work?

You cut a price, ran a campaign or launched an offer, and sales went up. Was that you, or would it have happened anyway? Counterfact tells you, in plain numbers.

Get a free check See how it works ↓
  • Answers in minutes, not weeks
  • No experiments or test groups needed
  • Says “we can't tell” when it can't
example

 

 

What happened What would have happened anyway

Why it matters

Stop guessing. Start knowing.

₹

Put money where it works

Find out which discounts, campaigns and launches actually brought in sales, and stop paying for the ones that didn't.

✕

Don't celebrate false wins

Sales often rise because of a festival, the season or the market. Counterfact separates your impact from everything that would have happened anyway.

✓

Decide with confidence

Get a clear answer with a likely range, or an honest “not enough evidence yet”, so you can back your decisions in any meeting.

How it works

One example, step by step.

01

Here are your weekly sales.

They already go up and down with seasons, festivals and the market.

02

You cut the price by 10%.

Afterwards, sales went up. But how much of that rise was because of your price cut?

03

The quick answer is misleading.

Comparing sales before and after says +25%. But Diwali came right after your change, and the whole market was growing. The quick answer gives your price cut credit for all of it.

04

Look at things you didn't change.

Sales of similar products and in other regions went up too. That shows what was happening anyway.

05

Counterfact works out what would have happened.

Using your history and those comparisons, it draws the sales you would most likely have had without the price cut.

06

The difference is what your change did.

Your price cut added about +6.5%, roughly a quarter of what the quick answer suggested. Now you know what it was really worth.

Your sales01 / 06
Your sales Things you didn't change Without your change

See the problem yourself

Why “before vs after” fools you.

Move the sliders. Choose how much your change really helped, add a festival or a growing market, and watch the quick answer get it wrong.

+10%
medium
+6%
normal
The quick answer says
–
What your change really did
–

Results

Tested on 1,200 real situations where we knew the answer.

We used real Indian and US data (state prices, stock indices, retail sales and jobs), added changes of a known size or none at all, and checked whether Counterfact found the truth.

9 in 10

Says “nothing happened” when nothing did

The usual tools see an effect that isn't there 4 times in 10, which leads to false wins and wasted budgets.

9 in 10

Its range includes the true answer

When it says “between +4% and +9%”, it means it. The usual tools' ranges hold only about 6 times in 10.

−16

Close to a famous study

On California's 1989 tobacco tax, it found a drop of about 16 cigarette packs per person a year. Researchers published about 19, well inside its range.

Built on an open model. Check the research version, its benchmarks and limits yourself on Hugging Face→

Who it's for

If you change things, it's for you.

₹

Brands & sellers

“Did the Diwali discount bring new sales, or just move them earlier?”

◎

Marketing teams

“Was the TV campaign worth what we spent?”

%

Lenders & fintechs

“Did the new approval rule lower defaults?”

§

Policy & research

“Did the new scheme actually reduce prices?”

More ways to use it

Measured, recorded, checkable.

Every result can carry a receipt in our public record, anchored daily on the Bitcoin blockchain, so nobody (us included) can change it later.

⏱

Change registry

Register a change before you launch. It's measured against exactly what you registered, so nobody can cherry-pick.
?

Claim Check

“This scheme raised X.” Public claims checked against the data, each with a receipt. Ask for a check, or prove a verdict wrong and get paid.
▦

Google Sheets

=CF_EFFECT() right next to your data. Free, two-minute install.
✓

Public record

Every market-board result and paid report, chained and anchored on Bitcoin, plus our public honesty score.
§

Policy tracker

Rate decisions, GST, tariffs and regulation, measured against sectors they shouldn't have touched.
♺

CSR impact

Evidence for Rule 8(3) impact assessments: what the programme changed, district by district.
⚙

For vendors

Put the engine in your analytics product: one API call, white-label output, verified receipts.
✎

Teaching kit

A free classroom lab with 20,000 worlds where the true effect is known.

Getting started

Three simple steps.

01

Tell us what changed

A price, a campaign, a launch or a policy, and the date it happened.

02

Share your numbers

A simple spreadsheet of weekly sales (or any number you track), plus a few things your change didn't affect, like other regions or products.

03

Get a clear answer

How much your change added or took away, a likely range, and a plain verdict: it worked, it didn't, or we can't tell yet.

Try it yourself

Have a spreadsheet? Check it now.

Upload a spreadsheet (CSV, XLS or XLSX) with your numbers week by week. Then just ask your question in plain words, or pick the details yourself. You get an answer you can download and share. Nothing is stored.

↧ Download a template
Load data to choose

Ask Counterfactask in plain words, after loading your file
Your answer will appear here.

Pricing

Start free. Pay when it's useful.

Free

₹0
  • Check any change on this page
  • Ask questions in plain words
  • Preview PDF (watermarked)
Try it now

Full report

₹499 / report
  • 10-page PDF: evidence grade, written analysis, period-by-period effects
  • 5 methods compared, fake-date test, backtest, sensitivity checks
  • Download it up to 5 times in 7 days
Check a change first
Best value

Pro

₹1,999 / month
  • Unlimited full reports for 31 days
  • Connect your Shopify store
  • Weekly “what worked” summary
  • API & Python package (2,000 calls)

Team

₹9,999 / month
  • Everything in Pro, for your whole team
  • More stores and data sources
  • We check your trickiest changes
Talk to us

Enterprise

Custom
  • Runs on your own servers
  • Your data never leaves
  • Security review and SLA
Talk to us
Already paid? Paste your access code to unlock full reports on this device.

Payments are processed securely by Razorpay. Prices include all charges shown at checkout.

Pro · Shopify

Connect your store. Get a weekly “what worked”.

Every Monday, Counterfact looks at your discounts and price changes and tells you which ones really brought in sales, and which didn't.

How to connect (2 minutes)

  1. In Shopify admin, go to Settings → Apps and sales channels → Develop apps, and create an app called “Counterfact”.
  2. Under Admin API access, allow read_orders, read_products and read_discounts, then install it.
  3. Copy the Admin API access token (it starts with shpat_) and paste it here. We store it encrypted and only ever read your data.

Needs Counterfact Pro.

Not on Shopify? Connect Razorpay

  1. In the Razorpay Dashboard, go to Account & Settings → API Keys and generate a key.
  2. Paste the key id and secret here. We store the secret encrypted and only read your captured payments.
  3. Register a change with your key id, and it's measured on your daily takings 2 weeks after the change.

Needs Counterfact Pro.

Questions

Good to know.

We're upfront about what Counterfact can and can't do, because a tool that always says “it worked” isn't worth much.

What do I need to get started?

Your numbers week by week (or day by day, or month by month), the date of your change, and if possible a few related numbers your change didn't affect, like sales in another region or of another product. A spreadsheet is enough.

How is this different from comparing before and after?

Before vs after mixes your impact with everything else that happened at the same time: festivals, seasons, the market. Counterfact separates them, so you only get credit for what you actually caused.

Do I need to run an experiment or A/B test?

No. Counterfact works on changes you've already made, including ones you could never test, like a national price change or a TV campaign.

What if it can't give a clear answer?

It tells you. If there isn't enough evidence, you'll see “we can't tell yet” rather than a confident guess. Often a few more weeks of data is enough.

When won't it work?

If something else big changed on the same day and there's no way to see it in your numbers, no tool can separate the two. It also needs some history: roughly twice as much “before” data as “after”.

Is my data safe?

The online tool doesn't store anything. For pilots we only need totals (like weekly sales), never personal data, and Counterfact can run entirely on your own servers.

Free check

Tell us about one change.
We'll tell you if it worked.

A price change, campaign, launch or policy. We'll get back to you within a day with what we need, and send your answer within 48 hours, free.

Prefer email? Write to rudraneel93@gmail.com