What we do

From trustworthy data to useful software

web_analytics

Web analytics

Understand what customers do and give your marketing team a measurement setup they can work with.

Deliverables, timing and details
  • GA4 and GTM, web and server
  • Consent Mode v2
  • Conversions API for Google Ads, Meta, TikTok

ProofAndy's agency experience includes a GA4 rollout across 20+ markets for a global beauty brand, inside three months.

data_science

Reporting & decision support

The maths behind a decision: how long a test must run, what a winner is worth, and which channel pays back.

Deliverables, timing and details
  • A/B test sizing and revenue projections
  • Funnel, retention and profitability analysis
  • BigQuery and Looker Studio dashboards
  • Python reporting with automated data checks

ProofAndy's work at an e-commerce experimentation platform includes sample-size calculations, revenue projections and customer-value reporting.

ai_engineering

AI & product engineering

Turn a workflow or product idea into an application, with the interface, integrations and operational tools it needs.

Deliverables, timing and details
  • AI-powered product features
  • Web applications and integrations
  • Reporting and document workflows
  • Reusable development and verification tools

ProofWe built Spotivibly: describe a musical mood, generate a Spotify playlist and convert playlists to YouTube.spotivibly.com

A repeatable path from idea to delivery

We reuse proven components and automate routine development work. Each project still gets a clear scope, a reviewable implementation and checks for the journeys your customers rely on.

  1. 01Clear scope
  2. 02Reusable components
  3. 03Accessible interfaces
  4. 04Data and access checks
  5. 05Documented handover

Products & tools

We build for ourselves, too.

Our own products and internal tools inform how we design, build and maintain software.

ProductLive

Spotivibly

Describe a mood to generate a Spotify playlist, or bring an existing playlist to YouTube.

  • AI
  • Spotify
  • YouTube
Explore product
ProductLive

QR Steady

Dynamic QR codes with editable destinations and scan reporting, so a printed code can keep working as your campaign changes.

  • QR
  • Analytics
  • Campaigns
Explore product
ProductLive

Who Sleeps Now

An interactive world map exploring where the world is asleep across time zones.

  • Data visualization
  • Maps
Explore product
ProductLive

FitMito

A personal coaching application designed to remember your context across conversations.

  • AI
  • Personalization
Explore product
ProductWork in progress

nVision Analytics

A server-first analytics platform designed to respect visitor consent state.

  • Analytics
  • Consent
  • Data infrastructure
ProductLive

AutoDev

A development framework for planning, building and checking software with coding agents and reusable workflows.

  • Developer tools
  • Automation
  • AI
Explore product
Internal toolLive

CV Studio

Our private workspace for editing and preparing CV documents.

  • Documents
  • Editing
Internal toolLive

Signature Studio

A private PDF workspace for placing your signature, reviewing documents and saving files on demand.

  • PDF
  • Documents
  • Private storage
ProductPlanned

nVision Builder

A planned website builder that assembles selected features from reusable modules. We are starting with the modules and their integration rules.

  • Modular software
  • Website generation

THE MEASUREMENT LAB

Follow the signal. Make a clearer decision.

A dashboard should answer a business question. Explore a fictional shop to see how sources become a shared view of performance.

Bring channel costs, sessions and orders into one agreed reporting model.

Deliverable: a source map, access checklist and consistent naming. Benefit: less manual reconciliation.

Synthetic demo · illustrative 30-day period · EUR

Revenue
€21,600
Channel cost
€4,500
Revenue / cost
4.8×
Order rate
1.6%
Paid search10,800
Paid social6,300
Email4,500

Order rate = orders / sessions. Revenue / cost is a descriptive ratio, not proof of incremental return. This example assigns each order to one channel; real attribution needs an agreed model. Costs here include only the listed channel costs.

From first question to useful reporting

Indicative delivery: 1–2 weeks for discovery and an initial dashboard; 3–6 weeks when tracking or integrations need implementation. We confirm scope after checking access, data quality and dependencies.

Track record

Experience across markets, products and teams

20+markets
Andy's agency work: a GA4 rollout for a global beauty brand, delivered inside three months.
3markets
Andy's agency work: a shared GA4 measurement model for a fashion marketplace's app and website.
2020founded
nVision combines direct client delivery, agency partnerships and its own digital products.
  • GA4
  • GTM web + server
  • BigQuery
  • Looker Studio
  • Segment
  • Google Ads
  • Meta
  • TikTok
  • LinkedIn
  • Microsoft Ads
  • Pinterest

Key events counted 162

Start with the reported count

Across four weeks, a lead-generation site reported 162 GA4 key events.

From real audits

Three things we found on live sites

Contact interactions counted twice

generate_lead + whatsapp_click, phone_click, form_submit

A summary event fired next to each event it summarised. GA4 reported 162 key events for 81 recorded contact interactions. The audit corrected the event count; it did not establish how many unique people became leads.

Consent granted before anyone clicked

gcs=G111 on page load, banner unanswered

Ad and analytics storage were reported as granted before the visitor made a choice. Consent Mode was installed, and it protected nothing.

Orders sent again on reload

purchase on order confirmation, page refreshed

Reloading the thank-you page sent the purchase a second time, so the same order was counted twice. A guard now sends each order once.

A useful place to start

What needs attention in your measurement?

Six questions. A practical next-step list based on your answers. No email needed.

Does every key report support a specific business decision?

For example: which landing page needs attention, and who can change it?

Are your conversion definitions written down?

A button click, a form attempt and a successful enquiry are different steps.

Have you checked when tracking is allowed to run?

Consider consent choices, collection settings and the information sent to each destination.

Have you tested success, failure and duplicate actions?

A working happy path is only one part of the measurement check.

Do reports show definitions, date ranges and data freshness?

Readers should be able to tell what was counted and how current it is.

Is someone responsible for checking measurement after changes?

Website releases, new forms and campaign launches can change what gets measured.

Packages

Fixed scope, priced in hours you can see

Every one-time price is the hours involved at €60 an hour. Pick one, or combine them into a single project.

Tracking audit

€480one-time · 8 h

  • GA4 property and GTM container review
  • Key events checked against what the site does
  • Consent Mode and ad platform signals
  • Written findings, ranked, with fixes

E-commerce tracking

€1,860one-time · 31 h

  • GA4 and GTM with the full e-commerce dataLayer, product to purchase
  • Consent Mode v2 banner and signals
  • Coordination with your developer
  • Documentation and a handover call

Add ad platforms at €300 each

Lead-gen tracking

€960one-time · 16 h

  • Forms, calls, WhatsApp and CTA clicks as key events
  • GA4 and GTM setup with content groups
  • Consent Mode v2 banner and signals
  • Documentation and a handover call

Add ad platforms at €300 each

Ad platform conversions

€300per platform · 5 h

  • Google Ads, Meta, TikTok, LinkedIn, Microsoft or Pinterest
  • Pixel plus standard events with values
  • QA with real test conversions

Server-side (Conversions API) is scoped separately

Looker Studio dashboard

€600one-time · 10 h

  • Sales, traffic and ad performance in one report
  • Cross-filtering, date ranges and previous-period comparison
  • GA4 and ad platform data together

Monthly care

€300per month · 6 h included

  • Fixes when the site or a platform changes
  • A monthly report
  • Data questions answered as saved Looker Studio reports

Minimum 3 months, extra hours at €54

Prices in EUR, excluding VAT. Payment: 50% at the start, 50% on delivery. Consent Mode v2 on its own is €240 (4 h). Data science and AI engineering work is scoped per project.

Ask for a package

How an engagement runs

Five steps, in this order

  1. 01 · first conversation

    Understand

    We review your problem, the people involved and what a useful result would look like.

  2. 02 · before we build

    Scope

    You get clear deliverables, dependencies and acceptance criteria, with a price and timeline for the agreed work.

  3. 03 · in reviewable steps

    Build

    We implement the tracking, reporting workflow or product, with working previews and regular feedback.

  4. 04 · before handover

    Verify

    We test the important journeys, data and access rules against the agreed acceptance criteria.

  5. 05 · optional

    Maintain

    A support plan for changes, reporting questions and the next improvements you want to make.

About

Measurement, models and machines

nVision Data is Andy's practice, running since 2020 and based in Romania. We work in English and Romanian, directly with e-commerce and lead-gen teams and alongside the agencies that run their ads.

Our work spans measurement, reporting automation and product engineering. Enterprise experience through agencies sits alongside direct nVision engagements and products we build ourselves.

Andy Nastasa
Andy NastasanVision Data

Ideas you can put to work

More clarity. Fewer assumptions.

Practical guides to measurement, reporting and evaluating AI advice.
Measurement

A form click is not a successful enquiry

Define form attempts, successful submissions and qualified leads separately, then test the cases that make a conversion report misleading.

Andy Nastasa
Reporting

Give every dashboard metric a definition

A practical metric brief: decision, population, formula, time window, freshness and owner. Make the number understandable before adding another chart.

Andy Nastasa

YOUR NEXT MOVE

Start with a good question.

Give us the context. We will use it to clarify the scope, dependencies and a sensible next step.

Where do you need help?

What would you like to improve or build?

Share your website, a reporting problem or a product idea. Tell us what success would look like and we'll help define the next step.