Setup pipeline
  1. Access
  2. Validate
  3. Fix
  4. Standardise
  5. Connect
the same order for every store

GA4 · BigQuery · Conversational analytics

Ecommerce Analytics That Counts What Actually Happened

We validate and fix your tracking first, then join every ad platform’s spend to revenue in your own BigQuery — and give you a standard dashboard set you can question in plain language.

About a week end to end · ~48h of build once accesses are in place
Server-side tracking
Your data sources
Google Ads Meta Ads TikTok Ads Shopify
store · ads · GA4 · BigQuery
Your standard dashboards
Acquisition
Data Shot
Funnel
Attribution
Cohorts
Runs in your own BigQuery One standard package, not a custom build Ask your data in plain language
Sources we connect
ShopifyWooCommerceGA4Google AdsMeta AdsTikTok AdsPinterest AdsMicrosoft AdsBigQueryLooker Studio

Everything you need to trust your numbers

One standard analytics system for ecommerce: validated tracking, all ad spend in one model, and answers you can get without waiting for an analyst.

Conversational analytics

Ask the data in plain language and keep digging with follow-up questions — no new ticket for every SQL cut.

Why did Campaign X’s ROAS decline last week?
Mobile ROAS fell 38%, desktop held. Two product pages drove most of the drop.
Break this down by device

Tracking & data validation

We start before the reporting: GA4, GTM, purchase and revenue events, pixels and UTMs get checked and fixed.

GA4 configurationvalid
purchase eventfixed
revenue valuevalid
duplicate eventsremoved
UTM usage · consentvalid

All ad spend in one model

Google, Meta, TikTok, Pinterest and Microsoft Ads costs joined to revenue — channel, campaign, ad set, ad.

Google Ads4.1× ROAS
Meta Ads2.7×
TikTok Ads1.4×

Attribution beyond last click

One attribution logic in BigQuery: first-touch, last-touch and assisted contribution side by side.

first-touch
assisted
last-touch

A weak last-click campaign can still be the first touch that starts the sale.

Your data stays yours

BigQuery and Looker run in your own infrastructure. No closed system, nothing to migrate out of later.

Your BigQueryyour cloud
Standard Looker packageready
Acquisition & funnel reportsstandard

How it works

From access to handover in about a week

The same nine steps run for every store. Most of the calendar time is you collecting accesses — once they are in place, the build itself takes roughly 48 hours.

  1. 01

    Access

    You grant a standard set: store admin, GA4, GTM, ad accounts, Google Cloud.

  2. 02

    Validation

    We check what your tracking actually sends: events, purchase, revenue, duplicates, consent.

  3. 03

    Fixing

    Broken events, misfiring tags and UTM gaps get repaired before anything is reported on.

  4. 04

    Integration

    Store, GA4 and every ad platform in scope are connected to one pipeline.

  5. 05

    BigQuery

    Data lands and is modelled in your Google Cloud project, under your billing.

  6. 06

    Dashboards

    The standard Looker package is deployed: acquisition, funnel, cohorts, trends.

  7. 07

    Conversational analytics

    You get a layer that answers questions in plain language and keeps the thread going.

  8. 08

    QA

    Numbers get reconciled against the store back office before anyone is asked to trust them.

  9. 09

    Handover

    Short videos and written instructions on reading the reports and asking the data questions.

Egor Dubinin, founder of Data Shot
One package, every store
the expertise lives in the product

Behind the standard

Built as a product, not as billable hours

An analytics freelancer can build you almost anything — that is exactly the problem. Requirements, implementation, changes, more analyst hours. Data Shot is the opposite bet: one validated pipeline, defined in advance, deployed the same way every time. What you are buying is the standard, not someone’s calendar.

Most dashboards fail long before the first chart: the purchase event fires twice and ad spend never joins to revenue. So the product starts where the data starts — and it starts the same way for everyone.

9 stepsidentical for every store
5 platformsad spend joined to revenue
Your cloudBigQuery and Looker, not ours
Egor Dubinin
Founder, Data Shot — product & implementation
Request setup

Pricing

Fix one thing, or standardise the whole stack

Most stores start with a single tracking problem. If the analytics underneath turns out to be worth rebuilding, the full setup is the next step — not a prerequisite.

Single fix
from $150

One defined technical problem, quoted and closed on its own.

Browse single fixes
Maintenance
$200–300 / month

Keeping the system running and telling you when it stops. Not unlimited support.

  • Standard system upkeep
  • Data pipeline monitoring
  • Tracking health monitoring
  • Standard integration monitoring
  • Continued conversational analytics access
  • Notifications when something technical breaks
What’s covered

What the standard package deliberately does not include

This is what keeps the setup fast and the price predictable. If many stores ask for the same thing, it becomes part of the product — not a one-off build.

  • Individually designed dashboards
  • Custom CRM integrations
  • Per-client attribution models
  • Arbitrary custom metrics
  • Non-standard sales funnels
  • Custom BI development

FAQ

Questions worth asking before you buy

How long does the setup take?

About a week end to end. Most of that is you collecting accesses — once all the correct accesses are in place, the technical deployment can be done in roughly 48 hours.

Do we get dashboards designed for our business?

No, and that is deliberate. Every store gets the same standard package: acquisition performance, ecommerce funnel, cohorts and trends. We do not copy your existing reports or design a bespoke dashboard per client — that is what keeps the setup a week rather than a quarter. If many stores ask for the same missing view, it becomes a standard feature of the product.

Who owns the data and the infrastructure?

You do. BigQuery and Looker run in your own Google Cloud project under your billing account. Your business data never moves into a closed system you would later have to migrate out of. Our access is collaborator-level and can be revoked without breaking your reports.

Does it show our actual profit?

Not in this version. Real profit needs COGS, shipping, payment fees and returns, and in most smaller ecommerce businesses that data either does not exist or lives in non-standard systems. We report on revenue, ad spend, ROAS, CAC/CPA and conversion rate — and we say so up front rather than quietly approximating.

How does the system decide what counts as bad performance?

You set your target ROAS. Above it is fine, below it needs attention. We are not shipping a benchmark or AI scoring engine at this stage, and we would rather give you a threshold you control than a number you cannot interrogate.

Which platforms do you connect?

Shopify and WooCommerce on the store side; Google Ads, Meta Ads, TikTok Ads, Pinterest Ads and Microsoft Ads on the advertising side; GA4 and GTM for tracking. Additional integrations get added to the product when demand for them repeats — not as one-off custom work.

How is this different from a self-service analytics SaaS?

A self-service tool says «connect your accounts and see your reports», and then inherits whatever your tracking is doing wrong. We start earlier: validate, fix, standardise, connect, and only then analyse. The setup fee buys a managed implementation, including verification that the data underneath is worth reporting on.

What exactly does monthly maintenance cover?

Keeping the standard system running: pipeline monitoring, tracking health monitoring, integration monitoring, continued access to conversational analytics and notifications when something breaks. It is not unlimited support. If Meta costs stop arriving because of your billing or an expired token, you fix it on your side; if a component of ours broke, fixing it is on us. If you change your store theme and break tracking, detecting that is covered — rebuilding it is separate work.

Start with the problem you already have

Tell us what is broken — a purchase event that never fires, conversions Google Ads never receives, a Meta pixel that double-counts. We scope that one fix. Whether the rest of your analytics is worth standardising is a separate conversation, after we have seen the data.