HOW WE WORK
Marketing data analysis in 4 steps
Most companies have too much data and too little information. This guide shows the method we use to go from scattered reports to data-driven decisions.
1. Data audit: trusting the numbers
Before analysing anything, we check that the data is right. The most common problems are conversions counted twice, sales without values, payment traffic (PayPal, card gateways) recorded as the source of the sale and data loss from missing Consent Mode.
Carrying out a data analysis first is, essentially, about reducing risk. How? Before investing in any campaign or strategy, data analysis shows what is already working, what is wasting budget and where the real opportunities are. It is the difference between an informed bet and a shot in the dark. Businesses that make data-driven decisions reduce their cost per acquisition by 38% on average and increase their conversion rate by 65%, because every action is directed to where there is real evidence of return.
Do you need access to my platforms to work? Yes, we need read access (and write access for technical configuration) to GA4, Google Tag Manager, ad platforms and, ideally, your CRM. All access is managed with the minimum permissions required and can be revoked at any time. We never share customer data with third parties.
- GA4: events, conversions, internal traffic filters and data retention.
- Ads: Google Ads and Meta conversions aligned with GA4.
- Store and CRM: orders, refunds and sales stages in sync.
2. Structuring and cleaning
We bring the sources into a common structure: the same definition of customer, conversion and channel on every platform. We standardise campaign names with UTM conventions, remove duplicates and connect marketing data to business data (margin, repeat purchases, customer lifetime value).
My data is confusing and I cannot understand what it is telling me. What could be wrong? It is more common than you might think. The most frequent problems are: GA4 misconfigured and counting duplicate visits, conversion events firing multiple times, bot traffic inflating metrics, and misattributed traffic sources that make "direct" look like the main channel when it is actually a tracking problem. Our audit delivers clean, reliable data ready for decision-making, usually in less than 48 hours.
3. Analysis and insights
With reliable data, the right questions get answers. We analyse the conversion funnel page by page, performance by channel and campaign, the audiences that buy more and those that only browse, and how customer value evolves over time (cohort and LTV analysis).
Every analysis ends with concrete, prioritised recommendations, not a 40-page report.
My target audience is too broad. How does data help refine segmentation? With real behavioural data, we can identify the audience segments with the highest purchase intent, highest LTV and lowest cost per conversion. Instead of speaking to "everyone aged 25 to 55", we start speaking to "users who visited the product page more than twice, came from Google Ads and dropped off at checkout". The difference in budget efficiency is substantial.
What is the difference between having data and having strategic insights? Data is numbers. Insights are answers to concrete business questions: "Which segment has the highest LTV?", "At which step of the funnel am I losing the most leads?", "Which channel has the lowest cost per customer acquired?". onOrbit Marketing does not deliver raw data reports; it delivers actionable answers with concrete recommendations for your next business decision.
4. Action and monitoring
Insights only matter if they change something. We follow the implementation of the recommendations (new campaigns, site changes, email segments) and measure the impact. Real-time dashboards have alerts for conversion drops, cost spikes or tracking failures.
Show 4 more topicsShow less
Dashboards: Looker Studio or Power BI?
Looker Studio is free, connects natively to GA4, Google Ads and Search Console and is enough for most SMEs. Power BI is better when you need to combine ERP, invoicing or several internal databases, or when the company already uses the Microsoft ecosystem.
In both, we design dashboards by role: management sees revenue, margin and cost per customer; the marketing team sees campaigns, channels and tests.
Attribution: which channel deserves credit for the sale
A customer may see an Instagram ad, search for the brand on Google days later and buy through an email. The attribution model decides which channel gets the credit. We compare GA4's data-driven model with ad platform attribution and incrementality tests, so you do not cut channels that are actually creating demand.
I am investing in Ads but the results are not convincing. Can the data explain why? In most cases, yes. Underperforming campaigns almost always have one of three causes that can be identified in the data: a target audience that is too broad (spending budget on people who will never buy), a landing page with friction that loses the user before conversion, or badly configured tracking that makes it look as if nothing converts when in reality it does, it just is not being measured. The onOrbit Marketing data audit identifies exactly what the problem is, with evidence, before recommending any changes.
Metrics every company should track
There is no universal dashboard, but some metrics almost every company should see every week:
- Cost per acquisition by channel and campaign.
- Conversion rate of the site and each funnel step.
- Average order value and repeat purchase rate.
- Customer lifetime value (LTV) compared with acquisition cost.
- Revenue by channel, separating brand and non-brand.
How much data analysis costs
onOrbit Marketing plans are Data Start (€390/month), with GA4 audit and setup, funnel analysis and a monthly report with recommendations; Data Intelligence (€790/month), which joins GA4, ads and CRM, with audience, LTV and cohort analysis, a real-time dashboard with alerts and a monthly meeting; and Data Enterprise (€1,490/month) for operations with several data sources and markets, with advanced Power BI or Looker Studio and a dedicated data manager.