Business Intelligence

Faster, Deeper, Profitable Insights

Two Thirds of Business Intelligence Projects Fail

Business Intelligence is a top priority for CIO’s, so failing to meet project goals, communication issues and excessively long delivery time frames are a real problem. We run our business intelligence projects in a fundamentally different way to avoid these pitfalls.


Business Intelligence projects often get into trouble with unclear objectives, an overly tight scope, lack of a solid business case, conflicting priorities, poor communication, and no clear vision on how the data will be monetised. This presents inter departmental "buy in" challenges and difficulties ensuring everybody is clear on the mission ahead. We do our best not to blame the technology and instead we work with you to deliver a warehouse, cube and visualisation capabilities that your organisation needs.


Our approach avoids choosing what data to include or not, but instead includes 100% of your data, saving time. Our build approach, which is automated can produce a self-building warehouse with tables, views, stored procedures and business functions in under 30 minutes. Once we have this, we then add in the custom elements your business needs. We also avoid creating hundreds of pre-defined reports but instead provide self service capabilities that can deliver every report output needed. We will data mine your data to tell you where the value is and how to monetise this to cover the cost of the project. Lastly, we build highly visual interactive dashboards to impress clients and your board. The net result is a cross business information solution that can be implemented more than six times faster than traditional methods in our experience, with significantly less debate on specification reducing communication issues. This typically results in completed solutions in under 3 months when a single consultant is assigned. This represents a huge cost saving on typical projects of this kind.

A True View of Business Performance

Effective Business Intelligence effectively answers questions in a way that delivers trustworthy, flexible and actionable results. We make sure all of your sources of data are linked and consistent across the organisation. This allows leadership to make more informed and strategic choices. We offer a self-service approach to information provision and empowerment that avoids the typical department bottlenecks.

Technical Methods

Our preferred software of choice is Microsoft SQL Server due to its market leading position, integration with office and SharePoint, and low cost of ownership. Both on premise and off premise services are available.

  • Less than 5% of code is hand written and hence is much less prone to errors and is significantly faster to produce.
  • A well structured naming convention is used to help keep everything clear and logically grouped.
  • Complex logic is broken down into small simple step views which execute faster, are easier to maintain, and run more robustly.
  • All logic is embedded in functions, with no logic held in stored procedures or SSIS.
  • SSIS is only used to transfer data and to orchestrate parallel running and alerting, which avoids embedding highly complex steps into hundreds of nested property pages. This also reduces the dependence on less ubiquitous SSIS skills when the solution is handed over, and more so on simple SQL skills enabling faster problem resolution.
  • SQL replication is used to ensure real time data is immediately available, which supports a 40 minute warehouse build time window and a 20 minute cube build time window.
  • Instead of spending the time defining which fields that should be included in the warehouse, every field is included by default in both the warehouse and a single cube. This is achieved with a special scheme design that allows several traditional cubes to be merged.
  • Once built, all metrics can then be defined and built within workshop sessions to speed up the creation process, as the building blocks are immediately at hand.
  • The single cube can then be used as the primary source for reporting and analytics, with the warehouse supporting bulk raw data style reports only. This saves having to build the metric fields into two locations and have two semantic models. It also takes advantage of the cubes ability to recut any metric along every dimension it is associated with. This can be difficult to achieve from the warehouse without having to embed logic directly into reports or a common view.
  • All data sources are mapped to an abstraction layer to enable the warehouse to be moved without impact and similarly there is an abstraction layer for the warehouse which incorporates user security at dimension record level, which is one step further than what SQL natively provides. In the cube, security is supported by roles and the use of Kerberos, although secure store accounts can also be used.
  • Each build is atomic, logged, and can be run with a full unit test in parallel, if needed.
  • Once everything is in place, self service becomes the primary method of creating BI assets (reports, dashboards, and data mining models) for both the reporting team and super users within the business. This significantly speeds up the report creation time as this is usually a drag and drop affair with virtually no more need for SQL queries.

Delivering BI quickly means all the benefits and requirements stay fresh, implementation costs are reduced, and you can react quickly to changes. Our solutions are designed to update hourly as a standard and therefore are specified to an appropriate hardware level.


Our automation system creates code that could take weeks to write, in just minutes. We scan the input data sources to learn about each field, its data quality, variety, density, etc. and create code-based on rule selections in a GUI. It also incorporates standard business functions which can be used to clean up data with the click of a mouse. These include functions for standardising first names which is useful for creating a single view of the customer to checking postcodes to ensure they are valid. The solution is modularised (shown below) to enable full automation or just for the parts of the system that are needed.

Modulised ETL Automation