Operational visibility
Service levels, fulfilment, capacity, downtime and exceptions in one place - including the operational detail that never reaches a finance report.
Data & Intelligence
Turn fragmented data into decisions.
Most companies do not have a data shortage. They have data in ten places, definitions that disagree, and reports that arrive too late to change anything.
ProjectData builds enterprise data platforms and the applications on top of them: connecting ERP, CRM, POS, WMS, e-commerce, finance and marketing systems into one structured layer, then turning that layer into dashboards, analytics, forecasts and decision-support tools the business actually uses.
The situation
Enough data. Not enough information.
Your ERP knows what was ordered and what it cost. Your CRM knows who the customer is and what was promised. Your POS or e-commerce platform knows what actually sold. Marketing platforms know what was spent to acquire the customer. Finance knows the margin - a month later.
Each system is correct within its own boundary. The business question almost always crosses those boundaries, so somebody exports to a spreadsheet, reconciles it by hand, and the answer arrives with a delay and a caveat.
That manual reconciliation is the cost. It is also the reason management decisions get made on partial information.
Featured use case
From marketing spend to actual revenue.
The most common version of this problem - and usually the one with the clearest commercial payback.
- Marketing spend
- Leads / customers
- CRM
- Sales
- ERP / POS
- Revenue & margin
- Better decisions
What usually breaks
Marketing platforms report conversions using their own attribution windows and their own definition of a customer. The CRM holds a different identity. The ERP or POS holds the transaction that actually matters. Nothing joins them, so campaign performance is judged on leads and clicks while the commercial team works from invoiced revenue.
What we build
- Integrations into advertising platforms, CRM, e-commerce, ERP and POS
- A customer and order identity model that survives across systems
- An explicit, documented attribution approach - agreed with the business, not assumed
- Pipelines with a refresh frequency matched to the decisions being made
- Reporting on revenue, margin, repeat purchase and customer value by channel, campaign and segment
What becomes possible
Understanding revenue and margin by channel and campaign rather than platform-reported conversions. Combining marketing and commercial data in one view. Removing the reporting delay between spend and outcome. Feeding real business results back into budget decisions instead of proxy metrics.
Beyond marketing
The same layer answers other questions.
Once systems are connected and a shared model exists, the next problems are cheaper to solve than the first one.
Sales intelligence
Pipeline quality, win rates, discounting, customer profitability and account behaviour joined to what was actually invoiced and delivered.
Inventory & supply
Stock positions, movement, ageing and supplier performance across locations and systems, at a frequency the operation can act on.
Forecasting
Demand, revenue and capacity forecasts built on the company's own history, with the assumptions visible rather than buried in a model.
Management reporting
One reliable weekly or daily view for the leadership team, reconciled against finance so the numbers survive scrutiny.
Custom data applications
Where a dashboard is not enough: pricing tools, allocation tools, planning tools and decision-support applications built on the same data layer.
Capabilities
What we design and build.
Delivered from architecture through to the application people open every morning.
-
Data architecture
The model, storage strategy and boundaries: what is the source of truth for each concept, where it is transformed, and who owns it. Designed to be operated by your team, not only by us.
-
Integration & pipelines
ERP, CRM, WMS, POS, e-commerce, finance, marketing and internal systems connected through APIs, databases, queues or files - with monitoring, retries and failure handling, because integrations break in production.
-
Unified data layer
A consistent set of definitions for customer, order, product, channel, cost and margin, so the same question returns the same answer regardless of who asks it.
-
Real-time and near-real-time data
Event-driven or high-frequency pipelines where the decision genuinely needs them, and scheduled batches where it does not.
-
Business intelligence & analytics
Management dashboards, operational analytics and self-service reporting - built around the decisions being made rather than around the available fields.
-
Forecasting & decision support
Statistical and machine-learning models where the data supports them, wrapped in tools that present a recommendation and its reasoning, not just a number.
-
Custom data applications
Interfaces for the people who have to act on the data: planners, buyers, commercial teams, operations managers.
Approach
The smallest architecture that solves the problem.
Data programmes fail by starting too big. We start with one question the business cannot currently answer, and make that answer trustworthy.
- 01
One decision first
Pick a decision that is currently made blind, and establish what it would take to inform it properly.
- 02
Map the sources
Identify which system genuinely owns each piece of information, and where definitions conflict today.
- 03
Build the thin slice
Integrate only what that decision requires, in production - not a prototype on an extract.
- 04
Reconcile
Compare the output against a source the business already trusts, and resolve the differences before anyone builds on it.
- 05
Extend
Add the next decision. The second one is faster, because the model and the pipelines already exist.
FAQ
Common questions about data projects.
What does an enterprise data platform actually do?
It collects data from the systems a company already runs - ERP, CRM, POS, WMS, e-commerce, finance, marketing tools, internal databases - and organises it into one consistent, queryable model. On top of that model you can build reporting, analytics, forecasting and data applications without going back to each source system every time.
The value is not the storage. It is that a question like “what was our margin by channel last week” has one answer instead of four.
Can ProjectData connect ERP, CRM, WMS and marketing data?
Yes. Connecting systems that were never designed to talk to each other is the most common work we do. We integrate through documented APIs where they exist, and through database access, exports, message queues, file drops or middleware where they do not.
The technical connection is rarely the hard part. Agreeing what a “customer”, an “order” or a “channel” means across systems usually is, and that is part of the work.
How can marketing data be connected to actual sales?
By building the join between the systems that record acquisition and the systems that record revenue. In practice that means a shared customer and order identity, a defined attribution model, and a pipeline that brings campaign, CRM, and ERP or POS data into the same place on a predictable schedule.
Once that exists, revenue, margin and repeat purchase can be reported by campaign, channel and segment rather than inferred from platform-reported conversions.
Does ProjectData replace an existing ERP or CRM?
Normally, no. We build the intelligent layer on top of the systems you already have. Replacing a core system is a large, risky programme that is only justified in specific circumstances, and it is rarely what a visibility or workflow problem actually requires.
Where a system is genuinely the constraint, we will say so - and scope the smallest change that removes the constraint.
Can business data be available in near real time?
Often, yes - but it is a design decision, not a default. Some decisions need data within seconds (stock availability, fraud signals, operational alerts). Many management questions are answered perfectly well by an hourly or daily refresh, at a fraction of the cost and complexity.
We size refresh frequency to the decision being made, and to what the source systems can sustain.
What if our data quality is poor?
That is the normal starting position, and it is usually visible only once systems are connected. We treat data quality as part of the build: validation at ingestion, explicit handling of duplicates and missing values, and reconciliation against a source the business already trusts - typically finance.
A number nobody trusts is worse than no number at all, so reconciliation comes before dashboards.
What decision is your business currently making blind?
Tell us the question you cannot answer reliably today. We'll tell you which systems hold the answer and what it would take to join them.