Selected work · Case studies
Just as no two spreadsheets are the same, no two databases are the same.
What does good database
design look like?
We’re limited in what we can show – much of what we build holds confidential data. But by taking a proper look at three of the projects we’ve worked on, you can get a real feel for what’s possible.
Where every project begins
Whether it’s tracking music or students, sales or test tubes, all of our databases begin life the same way – with a meeting, and a single, deliberately open question:
In an ideal world, what would your database do for you?
Don’t think about your existing database. Think about what you’d like if absolutely anything were possible. We can always say “No – that’s not viable” – but sometimes you’d be surprised what can be done.
One of the world’s top universities, and a complex set of requirements.
When Imperial College London came to us, it was clear nothing “off the shelf” would ever work. The needs of Chemical Engineering were so bespoke, so variable, so full of exceptions and variations, that we knew we needed to create a hugely flexible solution.

The challenge
A single database that tracks – and relates – all of the following:
The database tracks
- Students
- Lecturers
- Degree programmes
- Courses
- Exam questions
- Aggregate marks / year
- Exams
- Marks
- Study-abroad rules
- Coursework
- Absences
- …and many more
Each course can feature both exams and coursework – and the exams might carry higher (or lower) weighting than the coursework in deciding whether a student passes.
Courses can be “must pass” or entirely optional; one might count for 20% of the year, another 5% – and the same course can count differently for students on different degrees.
Each year counts differently from the last – except the years that don’t count that way at all, like a year spent working in industry. And reports must be produced in highly specific formats for governing bodies – formats that change year by year.
The solution
Yes, this was – and is – a technically challenging project. But the truth of it is that success is as much about the relationship as it is about the database.
It’s a relationship where the client can call to say “we’re in the middle of exam season and the rules have just changed” – and know we’ll move heaven and earth to make it work, at very short notice.
They trust that we’ll keep working until we get it right, and be honest when we’re struggling to understand the latest rule change. Often it turns out we’re not alone in battling to “get it” – and we work it out together.
It’s why one department recommended us to the next:
- Chem Eng.
- Mech Eng.
- Aeronautics
- Electrical & Electronic Eng.
A decade on, we’re a trusted partner for the College – building large databases, writing small macros to make Excel tasks easier, and providing development and training throughout.
What it looks like
A few sample screens from the Exams database – much of the real data is confidential.
Ever bought or sold a house? Imagine doing it dozens of times a month.
Imagine how you’d feel if an agent got the details wrong about your house – forgot the price, or when you listed it. That’s why data is critical to an estate agent. When a top London agent came to us for a system to suit their specific needs, we were delighted to help.

The problem
Selling a property involves numerous moving parts. The database had to track them all:
Managed by the system
- Properties
- Commissions
- Multiple offices
- Sale dates
- Staff & negotiators
- Cancelled sales
- Contract exchange
- Multiple sales / property
- Monthly reporting
- Pricing
- Financial records
- Acquisition sources
The solution
The key was to understand exactly what role the database would play in the business. This wasn’t a system for customer-facing information – no descriptions, no photos of beautifully decorated rooms, no viewing logs. It’s a business tool, built to answer business-critical questions:
- What’s the average sale price per square foot of space?
- How many sales per month, per office?
- How many sales are aborted – and what percentage is that of the whole?
- How many sales does each negotiator achieve – and what commission is due?
Understanding how the financials are calculated was a challenging task, and one we gave real time to. Making that information easy to see – with clear red / amber / green status on the figures that matter – was central to the requirement.
- On track
- Watch
- Unpaid / overdue
What it looks like
An independent forensic laboratory, where accuracy is everything.
We’ve worked with Gradko for over a decade, starting with some highly customised training on database design. In this field there are myriad regulations to follow, everything is audited regularly, and accuracy is critical – and must be provable.
The challenge
Every test tube must be tracked through its entire life – and at any moment, you must be able to say exactly where it is, and everywhere it has been.
The journey of a single tube
- 01Order & dispatchTubes sent out to the client, logged against the order.
- 02Receipt back into the labWho received it, when – captured by barcode at the highest throughput.
- 03AnalysisWhich machine, which requirements, which analyst – results captured from GCMS.
- 04Double-checkingEvery result verified and signed off by a second pair of eyes.
- 05Results to clientSent by whom, to whom, when – with the PDF stored and locatable.
- 06ReconditioningBack into the library, ready for its next use – and counted.
And from a business perspective, it’s imperative to see where the backlogs are – tubes awaiting reconditioning, or sent out but not yet received. When thousands of tubes move each day, that visibility is everything.
The solution
We’ve worked very closely with Gradko, adjusting the system as the business has changed. We worked with lab staff to capture output from GCMS analysis machines in the format customers require; with the despatch team to build a barcode-reading system matched exactly to their process, for the highest possible throughput; and with management to give a clear workflow overview, with automated alerts the moment a bottleneck appears.
What it looks like
Confidentiality limits what we can show – but these give a good flavour.

Incoming tubes and the analysis each needs – so supervisors can assign work. 
Received tubes with exposure detail, logged against locations for pollution mapping. 
Customers and orders – with fast-track and overdue invoices clearly flagged. 
The tube library – type, location, uses so far, and time to reconditioning.
Your project could be next
Contact us to find your perfect database.
Every one of these started with the same free 30-minute conversation – and one open question about what you'd want if anything were possible.







