How to Get Funding for your DATA QUALITY initiatives
W H I T E PA P E
R
Approach for
getting Data Quality initiative funding from executives
Santosh Kumar Dubey
Data Quality and
Master Data Management Enablement Specialist
Contents
Introduction
I have implemented several projects in MDM (Master Data
Management) and DQ (Data Quality) space. I have created solutions which have
saved millions of dollars for customers in their organization by avoiding
Regulatory Fines, Trusted data for cross-sell up-sell, efficient supply chain
and many more. The story of Data Quality projects looks very rosy but there are
many hurdles to get it in to this stage. Presenting a Data Quality initiative
to CEO or Executives for getting the funding is One of the key problem most of IT
heads are worried about. I have tried to address the approach for getting buy
in from company executives in such initiatives.
The objective to have consistent, accurate, complete data may
open Pandora’s box in today’s Big data world. Program managers must be clear on
objectives and do their initial assessment of data. The boundaries of data
quality project roadmap must be clearly carved out. In case of initiative
funding demand, it must be clear to the project leads what they are trying to
do they are looking for just Data quality project or are trying to solve
everything that organization is dealing with. The initiatives like Big Data
Management, Master Data Management, Data Lakes, Data Governance can have data
quality as one of their objectives. In case we are not starting these big
projects to solve our Data issues still independently DQ can solve issues like
wrong addresses, wrong phone numbers, wrong email id’s. The current era of
services provided by cloud, services provided by individual organizations to
correct or complete the data can help us in getting more out of DQ initiatives
without taking longer than expected time.
1.
Understanding
Data Quality
I would recommend to have simple assessment done for data before
even presenting it to broader audience within organization. In data quality
improvement journey, there are well defined definitions to identify the quality
of existing data. There are well defined 6 data quality standards to check
status of your data
1. Completeness
2. Conformity
3. Consistency
4. Accuracy
5. Duplication
6. Integrity
I.
Completeness
This is one of the key character of data, in this we try to get
the answers in terms which can be relevant to business. Is the address, email,
phone data populated, if not what is the impact it is having on business
objective. One of the key aspects or situations can be that we are not able to
deliver products to customer because we don’t have proper address of the
customer. One more check for data is whether address, email, phone data currently
available with us can be used as in state or not. This can have complex
outcomes as we might be thinking that data is present with us but due to
corrupt data it cannot be used for any purpose. If address information is
filled, may be the pin code is not valid or the state information is not
properly populated in the address which makes address information irrelevant.
II.
Conformity
This is the character of data where we try to look at data whether
it is matching the expected specification
We can check whether data values meet specified formats? If so, do all the values
conform to those formats? One of the key takeaways from this can be data
integration in IT terms or information sharing for the Business like the
downstream systems may not be giving proper results when you search all
customers staying in Bangalore where field stores different values like BLR,
BANGALORE, BENGALURU.
III.
Consistency
This character of data gives us conflicting information in same
data fields for example we may have address line one storing the details about
building number or in some cases the location details. In city field, we might
be having state information stored.
IV.
Accuracy
This is the character to check whether the values stored in data
base are having correct representation of data or not, for example the date
fields might be having date as 00/00/0000 or 9999. The data must be correct and
make sense. There are chances where due to some constraints put on UI’s users
have entered wrong data to bypass the check for example in date of birth field
if we are looking at dates like 00/00/0000 or 11/11/1111 indicates data is not
correct there. Similarly, there can be issue if transaction dates are kept of
future and not the appropriate dates when data was created.
V.
Duplication
This is one of the problems in data which impacts business a
lot, there can be systems having millions or records and in that there are lot
many overlapping records for example if a company is growing by acquiring other
company in same region. The situation will lead multiple records who are
existing in same region. The problem for business is going to be sending
duplicate marketing material to all these overlapping customers in their
systems. There can be some procedures or practices utilizing this data will
give wrong results, like for compliant rules you will not be able to give exact
number of customers in your system.
VI.
Integrity
Data integrity is a fundamental component of data quality.
In its broadest use, “data integrity” refers to the accuracy and consistency
of data stored in a
database, data warehouse, data mart or other construct. In
some systems, you might be having different value for same things like in some
system you can have source system Id as identifier or it may store different value
in same field.
2.
Understanding
Business Objectives
In case we are trying to get
funding for our project we must be doing our ground work of understanding the
business and their current impacted areas, and how data quality initiative can
add value to those pain areas. I have found 4 key pointers most of
organizations are trying to achieve in broader perspective
1. Decrease Cost
2. Increase revenues
3. Increase confidence in brand
4. Decrease risks
Let me try to put these four
points to perspective
i.
Decrease Cost:
This is one of the key objective
for most of company in today’s competitive era. Every organization is trying to
cut down on costs. Think about a story told by executive where he or she is
reported of 30% materials send by marketing division to the wrong people and
returned as not delivered objects. Thinking about the cost involved and lost
opportunities due to poor data, CEO, CIO can be excited if we can showcase them
how we can reduce this number 30% to probably 10% which is less cost.
ii.
Increase revenues:
I am constantly reminded by
overall objective of most of the projects IT department is undertaking. The
number one objective or task for any director, CEO, boards is to increase the
revenue and give good return to investors. It can demand and push for
investment from program managers if we can show case the impact on increasing
revenue. If there is any proposal to make impact on marketing by having clean
and latest email id’s in systems can obviously not get rejected.
iii.
Increase confidence in brand:
One of the biggest threats
most of boards are constantly worried is to make themselves complaints to new
regulations. The impact of any fines put by authorities is not only going to
give the financial loss but also brand name is also at stake. If there are any
positioning that we can do from our project objective perspective to reduce
this from happening.
iv.
Decrease risks:
We are constantly looking for
loss of data in current age of data where we are having sophisticated attacks
and hack of data. The best thing to avoid customer loss due to any loss or
stealing of data by implementation of data quality projects where we can
identify the high-risk data and observing how we can avoid that.
3.
Data
quality project Mistakes done before start of project
I have seen lot may data quality projects not returning long run
objective but getting pushed as one of the technical initiatives that a
technology leader is trying to deliver. There are several misconceptions that
we are having in data quality space which
i.
No Current Data Quality statistics:
One of the biggest mistakes
we do in our data quality initiatives is to goto executives of project sponsors
without doing data quality basic assessment and have statics ready with us.
When I am saying data assessment I am not talking about millions of records, it
can be done with 100-200 records and produce some very basic statics like the
pin numbers are not 100% filled, or states are in different formats like some
locations it is mentioned KA and some places as KARNATAKA.
ii.
Loosing site of Opportunity and talking about
problems with data:
This comes as second mistakes
in my list as most of time I have seen IT stakeholders presenting the DQ
project as problem statement to business and showcasing them there are address information’s
missing, email ids without proper formats, wrong phone numbers. The best
approach to showcase such things is to look it from opportunity perspective. We
can talk about the benefit instead of talking about problems in data. The clear
statements to be made in front of IT stakeholders is ‘We can improve our
marketing campaign more effective if we have proper email id of customer, as
part of DQ initiative we can improve the email id trust in our systems’
iii.
Stakeholders are not involved:
This comes as third reason for failure of our Data
Quality projects. There are few data quality projects where we are not able to
align data quality projects with overall objective of organization. I have seen
projects where IT is running in silo without being involved in regular outcomes
of these initiatives.
iv.
No proper data Governance:
Data quality is a process and always ends up as a
single IT stakeholder driving it. This is the mistake which ends up giving very
less results as part of program. We must try to have data governance program
executed overall to improve the data quality which includes generation of data,
consumption of data, security, archiving all coming together. The value adds that
we as data quality expert can do is train people and make them aware about the
process required to fix the process. I have seen people getting confused with
data governance program and its objective biggest mistake made in such
initiatives to assume everyone is aware about the process, sometimes we need to
spend time in training people on the process.
v.
Try to resolve all by yourself:
Data quality is a long thread which is having processes
in its center and then all the supporting stuff to help get it executed. Never
think that data quality initiative can be executed by one department and specially
it must be clear to all the stakeholders from start that it is not going to be
a IT initiative. The data quality with its subset tasks is having lot of involvement
from various departments and Business heads to take decisions.
vi.
Seeing Data quality as data cleaning activity:
This is the common mistakes done by most of data
quality initiatives where we name our just try to avoid big picture and fix the
data quality initiative to just clean the data for organization. The process of
doing data quality for organization is not a simple process, it requires lot
more than just cleaning or standardizing the data. The data that is getting
generated, processed, consumed the entire process needs to get changed. I have
seen customers being very successful once they started to make Data quality as
process for their organization and data cleaning, data standardization just a part
of process. The important part for the journey of data is everyone in
organization from top to bottom is held accountable and it is not a single
activity done in silo.
vii.
Not aligning technology to business objective:
The objective of data quality projects is to solve
problems in data. The problem arises only when we start talking about solving a
data quality issue by cleaning or fixing some issues. The data quality initiative
is a process which needs to be executed in multiple phases. The process must be
established and technology must be used to improve the process. Technology must
be an enabler and we must do our assessment to capture various positive and
negative value adds from them in our overall process.
viii.
Try to own everything data quality talks about:
I have seen enterprise wide data initiatives start
too big, cost too much and deliver too little at the end because they try to
cover everything that comes under the umbrella of data quality. I would
recommend to start small and deliver something tangible to business as soon as
possible and once it is delivered you can use that success to deliver next
important objective.
So basically we try to identify the current
required problem statement, identify the reason, fix it if it comes under the
ambit of data quality project and celebrate the success with business inputs.
Continue the step from start with next objective.
4.
Steps
to align Business Executive objectives to data quality project
I would like to take myself at the customer end and think about
the benefits from a user perspective. I get frustrated with calls coming from
xyz bank asking to buy some products which I have already purchased from them,
or policies which I have already denied to buy, or product deliver at my home
address where I have asked to get it specifically delivered to my office
address and there are no calls to my mobile about status of delivery. These are
the same issues which impact business as well as customers.
In case I want to address some of these issues I must have better
understanding of our customers which means for IT Department ‘Data’. The reason
I am not getting updated information about my preferences because organization
is not having latest details about my phone or email or address. I need to
address such issues by improving the data quality issue in my organization.
There are two ways of approaching the data quality projects.
a.
Top Down Approach
b.
Bottom Up Approach
i.
Top Down Approach
Define Clear
business objectives and then define how data quality can solve those problems,
example we are not able to connect with our customers over phone for doing any
cross sell or upsell, or we are having close to 30% product not getting
delivered to mentioned address.
ii.
Bottom Up Approach
Define a solution
using data quality Example we are having 30% customers with wrong phone numbers
Executives are always looking for ROI on their investments the
clear defined problem statement in business terms make them realize the benefit
of a data quality project.
Top down approach is always welcomed by
executives as they can understand the language of business and foresee the
problems currently created in business earnings due to this. I have worked with
executives and in case we start talking to them about current gaps in addresses
or emails, it makes no sense to them there must be always a linking of such
gaps to business perspective.
The data if looked from IT perspective and communicated to
business will not have much meaning to them as if we look down it doesn’t
communicate to business what is the impact on business due to this.
|
Current gaps in
data (incomplete address)
|
|
We are not able
to communicate to our customers properly
|
|
There are 40% of
address in current set with this issue
|
Now if we can follow up the same assessment from a business
perspective where the same message is formed from the business perspective. We
try to convert our data assessment and give the appropriate value addition due
to this.
|
There are 40% of
address in current set with this issue
|
|
Current gaps in
data (incomplete address)
|
|
We are not able
to communicate to our customers properly
|
|
40% Loss in cross sell up sell
opportunities
|
Converting business objective to data quality project
objective
|
There are 40% of
address in current set with this issue
|
|
Current gaps in
data (incomplete address)
|
|
We are not able
to communicate to our customers properly
|
|
40% Loss in cross sell up sell
opportunities
|
5.
Conclusion
It is advisable to approach executives which clear vision of
organization and details how Data Quality project is going to solve the problem
for them. The good discussion must be starting pointing out the organization
objectives and how Data Quality is going to help. It can be clear vision of
less return rate of marketing campaign material, Great up sell and cross sell
opportunities, Overall improvement of Customer experience with organization
Comments
Post a Comment