Business analytics and insights
Numbers on their own do not tell you what to do. We work out what the data is saying about your sales, your processes and your customers, and what is worth changing because of it.
The gap between reporting and analysis
Reporting tells you what happened. Analysis tells you why, and what to do next. Most companies have solved the first and skipped the second, which is why a dashboard can be accurate, well designed and still not change a single decision.
The difference is not sophistication. It is the presence of a question. A report shows revenue by month for everyone who might want it. An analysis asks why revenue by month looks like that, and whether the answer is seasonality, a pricing change, one large customer, or a channel quietly deteriorating underneath a stable total.
That last case is the one worth naming. Aggregate numbers hide compensating movements all the time. A flat total often contains a growing segment and a shrinking one, and the flat line is the least informative thing about it.
Where we usually look first
Every business is different, but a few areas produce useful answers often enough to be worth checking early.
Seasonality against real change. Knowing the normal shape of your year prevents both panic and complacency. It also tells you when to spend on marketing and when to accept a quiet period rather than fight it.
Funnel and process bottlenecks. Somewhere between first contact and payment there is a stage losing more than the others. It is frequently not the stage people assume, and the cost of fixing it is usually lower than the cost of acquiring more traffic to compensate.
Channel and campaign comparison. Put on the same basis, the ranking of channels by cost per acquisition often differs from the ranking by customer value twelve months later. Budget usually follows the first ranking.
Customer concentration. Which customers or segments produce the margin, and how exposed you are if two of them leave.
Being honest about what data can prove
Some questions have clean answers in your data. Others do not, and saying so is more useful than producing a confident number that will not survive contact with reality.
Attribution is the standard example. With cookie restrictions, multi-device journeys and buying cycles measured in months, no model reconstructs the true path from first touch to purchase. Attribution models are useful as a consistent lens for comparing periods, and they are not measurements of cause. Where the budget at stake justifies it, a holdout test tells you more than any model will. It means turning a channel off in one region and watching what happens.
The same applies to small samples. A conversion difference between two segments of forty users each is usually noise. We would rather say that than build a strategy on it.
From findings to action
An analysis that ends in a document has done half the job. The other half is deciding what changes because of it, and that requires the people who own the process in the room.
We keep the output short and separate three things clearly: what the data shows, what it suggests, and what our recommendation is given both. Recommendations get ordered by expected effect against the effort to implement, because a list of twelve equally weighted improvements produces the same result as no list at all.
Each recommendation also comes with the measurement that will tell you whether it worked, agreed before the change is made. Deciding afterwards how to judge a change is how organisations convince themselves that everything they tried succeeded.
Analysis is a habit, not a project
The first analysis is usually the largest, because it doubles as an audit of what your data can and cannot support. After that, questions get cheaper to answer. The pipelines exist, the definitions are settled, and a new question is a few days of work rather than a project.
That is the point where analytics starts affecting how the company operates: not because there is more data, but because asking a question of it stopped being expensive.
What you get
Trend and seasonality analysis
We separate a real change in the business from the normal shape of your year, so a quiet August stops being treated as a crisis and a strong November stops being treated as a win.
Bottleneck identification
Where in the funnel or the process people drop out, how long each stage takes, and which stage is costing you the most in absolute terms rather than in percentage points.
Period and channel comparison
Campaigns, sales channels and traffic sources compared on the same basis, so the question of what works best and where budget is being wasted gets a direct answer.
Customer segmentation
Grouping by behaviour and value rather than by industry alone, which usually shows that a minority of customers generates most of the margin.
A written set of findings
A short document stating what the data shows, how confident we are in each conclusion, and what we could not determine from the data available.
Prioritised recommendations
Specific actions ranked by expected effect and by effort: what to strengthen, what to change, and what to stop doing.
How we work
- 01
Frame the question
We agree what decision the analysis is meant to inform. Analysis without a decision attached produces interesting charts and no consequences.
- 02
Check what the data can support
Before analysing we look at coverage, gaps and known quality problems. It is better to say a question cannot be answered reliably than to answer it from thin data.
- 03
Run the analysis
Cohorts, funnels, segment comparisons and time series, depending on the question. We test whether an apparent effect survives being sliced a different way.
- 04
Separate finding from opinion
The report distinguishes what the data shows, what it suggests, and what is our judgement based on experience. You should know which is which before acting.
- 05
Turn findings into actions
We work through the conclusions with the people who own the relevant processes and agree what to change, what to measure afterwards and by when.
Tools and technology
Where a solid open-source tool exists, we choose it over a closed one. No lock-in to a single vendor, and costs you can actually predict.
- PostgreSQL
- dbt
- Python
- pandas
- scikit-learn
- DuckDB
- Metabase
- Apache Superset
- PostHog
- Google BigQuery
- Looker Studio
- Google Analytics 4
Frequently asked questions
More services in this category
Data collection and organization
Your numbers live in five different tools and none of them quite agree. We bring them into one place, on a schedule, so a question about last month has exactly one answer.
Dashboards and reports
A dashboard is worth building when it changes what someone does on Monday morning. We design views around the decisions you make, not around every metric a tool can export.
Reporting automation
If someone spends the first day of every month rebuilding the same report, that work can run on a schedule instead. The report arrives, the numbers are current, and nobody assembles it by hand.
Read testimonials from companies that trusted us
Delivered well ahead of the deadline.
ZanReal's individual approach is impressive.
Knowledge and business intuition make them a valuable partner.
Quick solutions that reduced costs by 99%.
Have data, but no clear answer on what to change?
Message usDescribe the decision you are trying to make and the data you have. We will tell you whether it can answer the question, and what the analysis would look like.
Can't keep up with changes in AI world?
Let us do the heavy lifting. Every week we distill the most important AI developments into a focused 5-minute briefing — so you stay ahead without the noise.
Find out more