How to Create an Information Chain
by Dylan Jones, Editor
Understanding how data flows around your business to deliver value to your internal and external customers is critical to effective data quality management.
In this article we take a look at some simple techniques required to deliver effective information chain management.
Why is information chain management so important to your business?
Way back in 1992 I started my very first data quality assignment.
The problem we faced was that it was taking our team between 4-6 months to deliver a data “product” to new customers. This delay was stifling our ability to grow the business.
We looked at ways to improve and being inexperienced I looked initially to technology to answer our prayers.
I tuned, tweaked, coded and probed our existing information processing systems in an effort to get the end-to-end delivery times down.
I failed. We got a few days here and there shaved off but no big dents were made in the problem.
Around that time I became involved in the ISO-9001 certification process taking place in our company.
Being a fresh-faced graduate I viewed this as a bureaucratic headache that took me away from my day job. However, one of the simpler techniques we adopted had a massive reduction in our delivery timescales.
We implemented an incredibly simple data lineage process. We began identifying the flows of information from their earliest beginnings in the minor tributaries flowing into the business through to the final customer product we delivered.
This had a profound effect on the way we managed information and ran the team.
Instantly we could identify bottlenecks, wasted processing, defect hot-spots, cost-centres and areas ripe for immediate improvement.
A better term for data lineage is information chain management. The reasons we need information chain management are varied but the ultimate benefit is felt by the customer.
In my example above, we delivered better quality data more quickly, directly benefitting the business of our customers and creating a strong differentiator in the market.
Emphasize the Benefits to the Business
Understanding your information and tracking its movement across the organisation is not particularly difficult in itself but sustaining an information chain management process and getting buy-in from disparate business and technical groups can be a challenge.
It is vital you stress that information chain management is not just another administrative process but a key driver in creating a leaner, more efficient organisation.
One way I’ve succeeded at achieving this in the past is to link information chain management to performance improvement schemes in the business.
Lean, Six Sigma and ISO 9001, BS 5750 type initiatives can all utilise information chain management to great effect.
The Techniques and Tools of Information Chain Management
Where do we start?
This is a common challenge, for example, do we create new chains or model existing chains?
There is no hard and fast rule but most companies begin by modelling existing chains to identify areas for improvement.
You will almost certainly identify issues with your existing chains once the mapping and modelling process is complete so this is a great starting point for linking your data quality strategy to some positive gains.
A common example, the “Purchase Service” information flow
To explain the information chain process we will look at a typical information chain common to most companies, that of making a purchase and delivering a product.
In our example we will examine a customer purchasing a holiday for their family.
Buying a holiday can be done over the phone, on the web, via postal request or in person via an agent.
The first thing we realise is that our information chain actually starts outside of our organisation.
TIP: It is essential that you identify the flows into and out of your organisation as a matter of priority. You need to create a firewall around your business so that you understand the levels of data quality you require to deliver world-class service.
A data location is simply a node or staging post for your data.
In our example, the first data location is either the call-handling system, the web system, the postal receipt centre or an external travel agency.
We need to record each of these type of data locations uniquely and also record who their business and technical owners are for any future contact.
For example, we may find that there are frequent booking failures with the end hotel. Rather than trying to implement some cleansing of the data we must trace the fault back through the various data locations until we find the source which could be an incorrect call-handling process.
A business stage typically possesses one or more data stages and is simply a means of simplifying your model for reporting purposes.
For example, in a Telecoms fulfilment information chain there could be literally scores of data locations and channels so by defining a business stage that can link these different sub-components we create a simple but effective means of categorisation.
A data channel simply charts the movement of data from one node (or data location) to another.
In our example the postal receipt team may take a booking form and enter the details into a central booking system.
Here, the data has flowed from the customer, via the booking form to the postal team and onto the booking system.
By creating a unique identifier for this data flow we can identify any defects that occur along this channel and help us distinguish exactly which channel to focus our improvement efforts.
For example, if you repeatedly discover booking failures on this channel it may be worth re-designing the form to prevent mistakes.
Each data channel must have at least one data location associated with it, most flows will have two locations.
In our example, our order will be received by the mail team, then prepared on the system via a data entry clerk.
What other information should we record?
When creating information chains for the first time it is important to keep things simple. Focus on the key locations and flows initially then drill deeper if time and budget permit.
As you get more proficient you may also like to record additional information such as:
- Data storage type (eg. database format (eg. RDBMS, spreadsheet, paper based records)
- Ownership – business, technical, customer, data stewards etc.
- Criticality – some data locations are more important than others, this can be used as a simple prioritisation mechanism
- Description – what functions does the data channel or data location support?
- Data quality levels – depending on the maturity of your organisation you may have data quality metrics for each node available, very useful for identifying areas of improvement
TIP: David Loshin of Knowledge Integrity has written probably the most detailed account of information chains available in his book “Enterprise Knowledge Management: The Data Quality Approach“.
David has also written a useful article on the additional data elements he stores in his version of information chains and also how to incorporate information chains in business impact analysis (click here).
How Do We Store and Retrieve Information Chain Records?
The key to effective information chain management is to ensure the data you capture is regularly maintained so as to be useful to the wider business, development and data quality teams.
Whatever option you choose, ensure that the business has the appropriate controls in place to accurately govern your information chains, this is far more important than the storage and manipulation tool.
That said, Microsofy Visio is a great tool for managing the different information connections and flows as it is graphical yet can store data values.
If you are highly proficient with Visio you can even integrate your models into a back-end database but this may be excessive for beginners to the process.
Microsoft Excel is a useful tool but a more interrogative tool such as an RDBMS (eg. MS-Access/Oracle etc.) is preferable as it allows more in-depth queries plus it allows you to link your information chain model to your wider data quality environment.
An alternative tool I have personally used for information chain management are the new breed of online application builders such as Zoho Creator instead.
How do we Enforce Control Outside of our Organisation?
The objective of information chain analysis is to document, measure and improve your information chains.
To do this you need agreements not just internally but also externally, with suppliers and even customers.
The key here is to pull, don’t push. If you beat a supplier up for lack of cooperation, you may find that when your new contract is written up they hammer you for additional charges.
Instead, get them to see the light by demonstrating the financial benefits you and your clients are witnessing. Explain that if they follow the same route they will pass those savings on internally to their business and obviously their clients.
We are drifting into data supplier management but this is an important aspect of information chain analysis as information chains are pivotal to supplier management and data quality governance.
Agree sustainable and attainable metrics
It is vital that you document and measure all information chains flowing into and out of your organisation.
Do not stretch your capabilities and those of your suppliers, go for attainable but important goals.
What Next for Your Information Chain Management Process?
So you have constructed an information chain with associated metadata but what value does this offer the business?
There is typically some business goal that means your information chain will be of value such as:
- Data migration or integration projects: Recording the flows of information into and out of systems is critical prior to decommissioning or commissioning a new system that will impact these flows
- Lean or Six Sigma projects: Lean in particular will benefit considerably from information chains when carrying out “value stream analysis”
- Data quality prioritisation: Information chain analysis will guide your analysts to the precise data locations that are critical to delivering customer services by tagging data quality measures to your information chains you start to create a far more effective means of prioritisation
- Process re-engineering: One of the main benefits of information chain analysis is that you identify ways to make the chain more efficient both to your own business and that of the customer (and indeed your data suppliers)
Summary of Information Chain Tips
- Information chains are a simple concept, easily taught but often difficult to enforce
- Focus on the key chains in your business as a matter of priority and link them directly to your data quality strategy
- If applied correctly they will have a tremendous impact on the efficiency and performance of your data quality initiatives
- Have you used information chains? What benefits did you accrue? Were they well received by the organisation?