Our colleagues in the telecommunications, cable, and Internet industries are all-too familiar with the “last mile” challenges of service delivery (i.e. the particular bandwidth, cost, and efficiency limitations that the service provider encounters when transitioning from the high-capacity central service bus to the end-point delivery mechanism, metaphorically the “last mile” of thetelephone line that leads to the customer’s house). But we are beginning to realise that we BI and data architects struggle with our own “last mile” problem with our data. Namely, while we often have strong processes to consume, digest, and store data, even in larger volumes and diverse types, we struggle to bring data (especially big data) home to our user communities within theorganisation to deliver on their needs.

For our data consumers, they do not care what effort it takes to bring them the data they need in the manner they require it, just as we are largely deaf to the excuses our phone, cable, or Internet providers offer when we experience an outage. That said, we are in somewhat of a unique situation with our data customers, in that we must create an effective partnership to understand the data needs they have and deliver from the often vast amounts of data at the organisation’s disposal to address them. But instead of a few dozen “products” to offer, we have essentially hundreds, if not thousands of data sources multiplied by numerous ways to prepare the data for consumption, and all of them must be accurate, robust, responsive, and extensible. For BI and data architects, this is our “last mile” problem.

For us, the last mile bottleneck is not one of cost or even really one of bandwidth (though some might reasonably argue that BI delivery on mobile devices endpoints must be matched to bandwidth-limited channels). Instead, our last mile problem more than any thing else is about usability. The big customer-facing challenge of BI is data usability.

By that we don’t mean to imply that we can’t make data usable. Of course, organisations have often a plethora of tools to manipulate data and draw data-based conclusions. Instead what we mean is that data needs to meet three criteria.

Data needs to be timely, accurate, and actionable. That is to say, we need to have data available at the time a decision must be made, it needs to be accurate (or correct) in order to draw conclusions based on facts as we know them, and it must somehow enable us to make a business decision, whether in response to a past or current situation or in anticipation of a future one.

A subtle insight here is that these criteria differ from user group to user group. While one might insist on data that is highly current, but accuracy is far less critical (say for real-time trending consumer sentiment), another group might be willing to wait for data, but it must be highly accurate against the company’s internal metadata (such as payroll reporting for tax purposes).

As we struggle to tackle the “last mile” data challenges of our organisations, the most important goal we’re striving toward is not to develop and adopt a monolithic standard by which we measure whether we’re delivering on the last mile, but instead a nuanced and multi-faceted measure of service delivery based on the customer’s real data needs.

Mr. Briggs has been active in the fields of Data Warehousing and Business Intelligence for the entirety of his 17-year career. He was responsible for the early adoption and promulgation of BI at one of the world’s largest consumer product companies and developed their initial BI competency centre. He has consulted with numerous other companies about effective BI practices. He holds a Master of Science degree in Computer Science from the University of Illinois at Urbana-Champaign and a Bachelor of Arts degree from Williams College (Mass)..
View Linkedin Profile->
Other Articles by Douglas->

Our colleagues in the telecommunications, cable, and Internet industries are all-too familiar with the “last mile” challenges of service delivery (i.e. the particular bandwidth, cost, and efficiency limitations that the service provider encounters when transitioning from the high-capacity central service bus to the end-point delivery mechanism, metaphorically the “last mile” of thetelephone line that leads to the customer’s house). But we are beginning to realise that we BI and data architects struggle with our own “last mile” problem with our data. Namely, while we often have strong processes to consume, digest, and store data, even in larger volumes and diverse types, we struggle to bring data (especially big data) home to our user communities within theorganisation to deliver on their needs.

For our data consumers, they do not care what effort it takes to bring them the data they need in the manner they require it, just as we are largely deaf to the excuses our phone, cable, or Internet providers offer when we experience an outage. That said, we are in somewhat of a unique situation with our data customers, in that we must create an effective partnership to understand the data needs they have and deliver from the often vast amounts of data at the organisation’s disposal to address them. But instead of a few dozen “products” to offer, we have essentially hundreds, if not thousands of data sources multiplied by numerous ways to prepare the data for consumption, and all of them must be accurate, robust, responsive, and extensible. For BI and data architects, this is our “last mile” problem.

For us, the last mile bottleneck is not one of cost or even really one of bandwidth (though some might reasonably argue that BI delivery on mobile devices endpoints must be matched to bandwidth-limited channels). Instead, our last mile problem more than any thing else is about usability. The big customer-facing challenge of BI is data usability.

By that we don’t mean to imply that we can’t make data usable. Of course, organisations have often a plethora of tools to manipulate data and draw data-based conclusions. Instead what we mean is that data needs to meet three criteria.

Data needs to be timely, accurate, and actionable. That is to say, we need to have data available at the time a decision must be made, it needs to be accurate (or correct) in order to draw conclusions based on facts as we know them, and it must somehow enable us to make a business decision, whether in response to a past or current situation or in anticipation of a future one.

A subtle insight here is that these criteria differ from user group to user group. While one might insist on data that is highly current, but accuracy is far less critical (say for real-time trending consumer sentiment), another group might be willing to wait for data, but it must be highly accurate against the company’s internal metadata (such as payroll reporting for tax purposes).

As we struggle to tackle the “last mile” data challenges of our organisations, the most important goal we’re striving toward is not to develop and adopt a monolithic standard by which we measure whether we’re delivering on the last mile, but instead a nuanced and multi-faceted measure of service delivery based on the customer’s real data needs.

Mr. Briggs has been active in the fields of Data Warehousing and Business Intelligence for the entirety of his 17-year career. He was responsible for the early adoption and promulgation of BI at one of the world’s largest consumer product companies and developed their initial BI competency centre. He has consulted with numerous other companies about effective BI practices. He holds a Master of Science degree in Computer Science from the University of Illinois at Urbana-Champaign and a Bachelor of Arts degree from Williams College (Mass)..
View Linkedin Profile->
Other Articles by Douglas->