Armin Sanjari
Managing Director, mathematics and analytics
Founded the agency IT-WINGS in 2016 and still runs it. It serves well-known clients in the automotive, consumer goods, pharma and mechanical engineering industries.
Data analytics consulting with AI.Numbers your team trusts.
Everything else in marketing rests on clean tracking and reliable data. We bring your sources together and build dashboards and analyses you can base decisions on.
Sources
Website, shop, ad accounts, CRM
Capture
Check consent, server-side tracking
Combine
Data pipelines into your warehouse
Analyze
Dashboards and AI analysis
A decision your team trusts
What does data analytics consulting cover?
Data analytics consulting makes sure you know where your inquiries and revenue come from. It starts with tracking on your website and shop, connected to your ad accounts, with visitor consent handled correctly.
Data engineering goes a step further. Data pipelines bring your ad accounts, analytics, CRM and shop together in one place, so reports no longer depend on exports and standalone Excel sheets.
The two go hand in hand. Without clean data, any automation runs on the wrong signals.
Spread across dozens of systems
Nobody has brought your CRM, shop, ad accounts and countless Excel sheets together yet.
Everything depends on one person
New reports take weeks. When that person is on vacation, nothing happens at all.
Gut feeling wins
The decision is made before the analysis is finished. The data would have known better.
What our data analytics work includes.
Tracking audit
We compare what Google Ads, Meta, Analytics and your CRM count. Then we find out why they differ.
Google Analytics 4 and Tag Manager
A clean setup with clear events and names that people will still understand a year from now.
Server-side tracking
Your own server receives the data and passes on only what is allowed and necessary. This makes measurement more stable and gives you more control.
Consent and GDPR
We link tracking to your consent banner and document what is collected. The legal assessment stays with your data protection officer.
Data pipelines
Data from ad accounts, analytics, CRM and shop flows automatically into a data warehouse. No more manual exports.
Dashboards
Looker Studio or another tool of your choice. A handful of numbers people actually look at, with cost per inquiry and results per channel.
Attribution
We show which channels work together. No model is perfect, so we also tell you how certain each conclusion is.
AI analysis
For questions no dashboard can answer. The AI looks for patterns, our specialists sanity-check every result.
Your data has answers. We find them.
AI data analysis goes far beyond marketing, from Excel spreadsheets to mountains of documents. You bring the question, we bring the method and the tools.
Sales & customers
Which customers are likely to churn, and which have room to grow?
Marketing & channels
Which channel actually drives revenue, and where are we burning budget?
Products & stock
Which items tie up capital, which will run out soon, and where is demand heading?
Documents & contracts
What do hundreds of contracts, minutes or reports say, and where do they contradict each other?
Service & feedback
What do customers really complain about? Tickets, emails and reviews, clustered automatically.
Processes & costs
Where do discrepancies, bottlenecks and duplicate work come from, and what can be automated?
How much potential does your data hold?
Four questions, instant assessment. No sign-up, and no data is stored.
Question 1 of 4
Where is your most important business data today?
One inquiry, four truths.
Every system counts differently. That’s the most common reason teams don’t trust their numbers. We clear up every discrepancy and decide which source counts for which question.
One inquiry, four truthsExample
- Inquiries in the CRM128What counts in the end
- Inquiries in Google Ads171counted twice
- Inquiries in Meta94Consent missing
- Inquiries in Analytics103Form not captured
CustomerRegion
Müller GmbHK-7F3ASouth
Anna SchmidtK-19C2North
Becker & Co. KGK-A04DWest
Jonas WagnerK-5E81East
The AI sees patterns, but no names.
If an AI analysis would touch personal data, we replace names with pseudonyms first. The AI sees patterns, not names.
Sample figures for illustration.
For us, data protection is architecture, not a footnote.
Pasting customer data into a public chat tool? Rightly off limits. Personal data is removed before the AI sees it. Everything after that runs in an isolated environment in the EU.
Your systems
Only the data needed for the question.
Pseudonymization
Names, emails and addresses are replaced.
AI + data scientist
Models in EU data centers, reviewed by us.
Answers instead of raw data
Only aggregated insights leave the environment.
Hosting in DE / EU
Processing exclusively in European data centers.
Pseudonymized
Personal identifiers are removed before the analysis.
No training
Your data isn’t used to train AI models.
Data processing agreement and deletion
Data processing under GDPR, deletion after the project ends.
From question to decision in four steps.
- 1
Sharpen the question
In an initial conversation, we work out which decision the data should support and what that takes.
- 2
Connect the data
By export or API, with a data processing agreement if needed. You keep control of every record.
- 3
Analyze and review
The AI prepares the data and looks for patterns. We sanity-check every result.
- 4
Hand over the results
Dashboard, report and clear recommendations. Also available as an analysis that keeps running automatically.
No data graveyards. Decisions.
Analysis report
Findings explained clearly, with charts and reasoning, even for people without a statistics background.
Prioritized actions
What you should tackle first, sorted by impact and effort.
Dashboard (optional)
Your key metrics live, so the insights don’t vanish after the workshop.
Automation (optional)
Turns a one-off analysis into a process that runs regularly.
We use this ourselves, every day and in production.
With Propspector, our AI takes minutes to review sensitive real estate documents such as the land register and condominium owners’ meeting minutes, GDPR-compliant and with processing in the EU. We use the same architecture for your data.
The AI does the grunt work. We make the decisions.
What the AI does
- Prepare data and find errors in datasets
- Detect patterns and anomalies
- Spot and report gaps in conversion tracking
- Draft queries and pipeline code
What our team does
- Clarify which question the analysis should answer
- Sanity-check models and results
- Decide which source counts for which question
- Make recommendations your team can act on
The people behind it.
More about us
David Otto
Managing Director, economics and growth
Has been scaling digital businesses since 2015, first at Festool, then for six years at Mercedes-Benz. There he led projects in digital products and global performance marketing.
Based in Stuttgart.Working across Germany.
The analysis itself runs in our secure environment, so where you’re based doesn’t matter. We meet in person wherever it helps the project: at the kick-off, in the results workshop or directly at your offices.
Stuttgart region, on site too
- Stuttgart
- Esslingen
- Ludwigsburg
- Böblingen
- Sindelfingen
- Waiblingen
- Reutlingen
- Heilbronn
All of Germany, remote and on site by arrangement
- Munich
- Frankfurt
- Karlsruhe
- Nuremberg
- Cologne
- Düsseldorf
- Hamburg
- Berlin
- Leipzig
MunichFrankfurtCologneHamburgBerlinLeipzigNurembergStuttgart
Frequently asked questions about Data Analytics and Data Engineering.
Related
Why do Google Ads and Analytics show different numbers?
They count by different rules. They handle attribution, time periods and missing consent differently. A small gap is normal. A large one points to errors in the setup.
What is server-side tracking?
With server-side tracking, the browser first sends the data to a server you control, which passes it on to Google, Meta and others. This makes measurement more stable and shows exactly what gets passed on. It doesn’t make up for missing consent.
What is the difference between data analytics and data engineering?
Data engineering builds the plumbing: the pipelines that collect data and combine it cleanly. Data analytics works through that data and answers questions. Without the first, the second is on shaky ground.
Do we need a data warehouse?
Not always. With only a few sources, a well-built dashboard is often enough. Once you’re combining several systems or want to keep data longer, it usually pays off.
Can Google Analytics 4 be used in a privacy-compliant way?
That depends on the setup, consent and contracts, and it’s a legal question for each individual case. We handle the clean technical setup and the documentation. The legal assessment is up to your data protection officer.
What does a tracking setup cost?
That depends on the number of sources, your consent banner and the depth you want. After the audit, you’ll get a specific quote.
What data do you need from us?
Only what’s really needed for your question. An export from your ERP, CRM or shop is often enough. For ongoing reporting, we connect the systems directly via API.
Does our data leave Germany or the EU?
No. Hosting and AI processing run in data centers in Germany or the EU. Personal data is pseudonymized beforehand.
Do you use ChatGPT or similar tools?
Not the public chat versions. We use AI models through business accounts in EU regions, embedded in our own infrastructure, with contracts that rule out training on your data.
What does an analysis cost?
That depends on the number of data sources, data quality, how deep the question goes and whether a dashboard or automation should follow. After the initial conversation, you’ll get a specific quote.
Do we need our own IT department?
No. We handle the integration, data preparation and operations. From you, we need a contact person who knows the question and the business.
Do you only work in the Stuttgart region?
No. We’re based in Stuttgart and serve clients across Germany. We mostly work remotely and meet on site when the project calls for it.
Who owns the results?
You do. Dashboards, pipelines and code are fully handed over to you, so you don’t stay dependent on us.
What question are you asking your data?
Tell us what you want to know. We will tell you whether and how your data can answer it. You’ll get a reply within 24 hours.
hello@kessel-und-code.de
Rotebühlplatz 11–15, 70178 Stuttgart