Am I getting the most Value of my SAS installation Dollars?

Managers think about SAS value particularly when it is renewal time!

In my many years as a SAS consultant, I have received this question several many times.  I remember one time in Michigan at a healthcare organization.  The executive mentioned he had a number of people across the organization approach him about issues related to the inefficient deployment and use of SAS. There was also a perception that the organization could or should be using SAS more effectively as a strategic advantage in certain situations.

This is a very common concern since organizations pay good money for their SAS software and wonder if they are squeezing the most value of those dollars.   Managers think about SAS value particularly when it is renewal time!  They want to know: what is SAS, how it is used across their organization as well as what are the challenges for them in fully leveraging SAS in their environment. 

Managers don’t even know what they have in terms of SAS software very often.  Does it sound weird? It is.  Especially considering the amount of money they are paying.  But if you are in this category don’t feel bad.  SAS licensing is so confusing and there are so many products and bundles that I have encountered this situation many times worldwide!

Would you agree the value of SAS for your organization comes from transforming data into actionable information using well prepared human resources?  If so, let’s look at seven areas where this potential value can be lost: inefficient data access, limited reporting and visualization, poor data cleansing, obsolete predictive analytics, incomplete SAS solutions, limited hardware use, and lack of governance.

INEFFICIENT DATA ACCESS

SAS has great capabilities as a 4GL to extract, transform, integrate and load data in both structured and unstructured data repositories (EDWs, Marts, RDBMs, data lakes).  I have found that SAS installations are poorly tuned and therefore subject to delays and contentions with other applications.  SAS installations can also be incomplete because managers don’t understand or have the perception that SAS does not operate on relatively new technologies such as Hadoop clusters.  It is also common that organizations have stopped at the SAS Base level use and have not moved into SAS Enterprise Guide, a newer interface.  Another possibility, they may have been using SAS EG but are not using SAS Data Integration suite of products.  This makes the overall SAS data management operation obsolete and clumsy.

LIMITED REPORTING AND VISUALIZATION

SAS BI has the capability to produce reports, dashboards, and mobile reports.  Many times, nobody knows how to use SAS BI software.  I believe, one reason is the traditional SAS BI products (SAS Business Intelligence components including Web Report Studio, Information Map Studio, SAS Information Delivery Portal and the SAS Add-In for Microsoft Office) were developed as modules and therefore managers have a hard time trying to figure out how these modules come together or what they do. The new generation of SAS BI products, like SAS Visual Analytics, are bundled a bit better but they still have technical requirements like the SAS LASR that makes these products harder to implement when compared with competitive products like Tableau, BO and Microsoft Power BI. 

 POOR DATA CLEANSING

If your data sources are not reliable then it is virtually impossible to get the value of your SAS installation. When I met with the board of directors of a large mobile operator in Africa, they told me they would regularly have different version of the “truth” depending on who was reporting.  This had created a lack of trust in the data and in the mechanisms to bring the data from source to consumption.  This is critical. Your SAS Data Management software (Dataflux) should have good matching, survivorship and cleansing rules to deliver quality data downstream.  Otherwise, data consistency and value are compromised. This situation hinders fact-based decision making – a significant value generator!

OBSOLETE PREDICTIVE ANALYTICS

Predictive analytics is the area where typically more value is created in the SAS installation.  However, I have found that scientists without training or exposure to best practices become stagnant in SAS STAT and never evolve into using more automated tools like SAS Enterprise Miner or solutions like SAS Model Manager that deliver additional capabilities to produce more and better models. I am not even mentioning SAS Viya, the new, high performance analytics architecture unveiled recently by SAS at its user group conference in Las Vegas. A Spanish bank can leverage efficient modeling to run more than one hundred marketing campaigns per month using SAS Customer Intelligence.  Can your team do the same?  That’s one typical way to leverage SAS software to generate exponential value.

 INCOMPLETE SAS SOLUTIONS

What about my SAS solution?  Most SAS solutions are a collection of base modules plus specialized modules like SAS EM as well as some customized modules according to the application domain i.e. healthcare fraud framework, gaming patron value optimization, service part optimization, retail merchandizing, etc.  These solutions have been created organically.  Some times they perform poorly because they are either being used partially in a modular fashion or the solution has not been integrated properly. 

POOR HARDWARE PERFORMANCE

Many organizations operate in silos supported by departmental computers.  It could be government vs. commercial.  Or finance, marketing, customer support, etc.  These SAS computing silos provide little strategic advantage for licensing economies of scale and make it difficult to improve hardware performance.  Look beyond departmental computing to unleash the value of your existing hardware using new technologies like SAS GRID that bring the promise to provide additional performance and throughput.   The SAS Grid Manager components make grid-enabled SAS applications available to a variety of SAS customers. SAS DI Studio and SAS Enterprise Miner have enhanced their integrated development environments (IDE) to provide grid automation by automatically generating SAS code for applications that are enabled to execute in a grid.

LACK OF GOVERNANCE

Look around and beyond SAS.  Perhaps best practices sharing do not exist and there is a lack of governance.  Develop processes and procedures to unleash the power of the SAS team. Otherwise, the information delivery process comes to a screeching halt.  For instance, is there an internal data dictionary or business glossary? Do you have data stewards?  Do you have too many data stewards?  You may find that there are some issues in transforming information which are "too difficult to correct".  This situation handicaps the ability to derive value from SAS.

WHAT CAN YOU DO?

• Review your SAS license and figure out what you have and what you should use.
• Conduct and assessment to break down business requirements and determine how best to leverage the existing SAS tools within your environment to meet project objectives.  Eliminate tools not needed.
• Create a SAS toolset.  Design and develop SAS standard macros, small applications, and other utilities to expedite SAS programming activities and data analysis/review
• Perform SAS programming quality control checks against source data and document all data issues
• Evaluate and improve existing SAS analytic models
• Evaluate cost of using SAS GRID
• If appropriate, build and operationalize analytical models within the SAS Grid environment, leveraging tools and data from various sources (Teradata, Hadoop, Excel files, MySQL, Oracle etc.)
• Review your SAS solution for completeness and integration
• Ensure all SAS programming activities and processes performed are conducted according to standard operating procedures and good programming practices
• Create a SAS group to provide guidance to team members on SAS solutions, best practices, and standards

If you need help evaluating your situation, please reach out.  We have in-depth knowledge of SAS software as we were once SAS employees. We know various best practices for configuring the software to best suit needs, have experience in SAS solutions, and can lead re-engineering sessions. We provide SAS technical expertise in building, integrations, testing and training of SAS applications.

Hope these ideas help.

Best regards.  Al 

al.cordoba@qlx.com

954-980-5992

Seven ridiculously simple and practical soft skills to become a truly fascinating data scientist!

There is a fascination test link at the end of this post.   Take it and you’ll be surprised with yourself!

Many people agree that being a data scientist is not only very fashionable but also challenging.  Of course, the definition of “what a data scientist is” has not been entirely decided.  Some people think a data scientist is some kind of a technically gifted “geek” ready to use a magic wand and transform legacy systems into a fascinating set of perfectly usable and well-coordinated set of new technologies like big data, in-memory application, nosql, cloud, etc. that will bring instant value to the organization. Great!

However, there is a catch.  You, as a data scientist, will need soft skills beyond the technical skills to make these dreams come true…. let’s say fascinating!  Those soft skills include: setting expectations properly, understanding the environment both technically and politically, creating a compelling vision, getting alignment between different groups, influencing your audience, building trust, and educating people with different backgrounds on new technologies.

1. Setting expectations

Are the expectations clear? Are you planning to involve others in setting up the expectations?  Like in a “What do you propose model”?  Or are you directing others based on a decision previously made.  Those are important considerations when setting up expectations. Core values should be used such as: providing valid and complete information, and allowing for free and informed choices.  It is absolutely unfair to demand some sort of team performance when expectation have not been made absolutely clear.

What gets defined and measured gets done.  Thinks about your objectives.  What do you want to achieve?  How will you measure success? What kind of indicators could you use? How will you know the objective is reached? Who do you want to fascinate? Think about the benefits.  What will it bring? What will it serve?  What value will be there that was not there before?

2. Are you listening?  – from active listening to fascination listening.

It is easy to stop listening when we think we know it all. The problem with this behavior is that other people are expecting you to listen.  Active listening means to listen with all your senses.  Fascination listening means to go beyond active listening into formulating smart questions with the intent to understand the human side. This the power of fascination listening.

It is critical to lead by questions. This way you may understand better the nature of the situation.   Key questions:

1.    What are successes? What did you achieve?

2.    How do you perform against your targets? – on, below, above

3.    What do you need to improve? Where can you stretch yourself?

4.    How can I support you?

 

3. Create vision – from solution to fascination value

Let’s redefine the challenge: how can you develop vision without consuming yourself, your team and your environment?

How to create vision? Consider some activities: Act with a sense of purpose. Feeling aligned, show yourself truly connected with ones-self and others. Radiate positive fascinating energy. Imagine and inspire a higher vision.  Reconcile performance and fulfilment in a sustainable way.

What can help you, and your team, to develop a truly connected, inspired and fascinated way of leading yourself, others, and your environment or project? Could it be “know how” (understanding), through state-of-the-art concepts, or “show how” (behaving), through workshops and role plays, or “be how” (living), through ‘experientials’ and symbols, or  a combination of all the above?

4. Get alignment –

As a data scientist, you may sense that the people around you are not fully engaged emotionally on the journey. They are not fascinated! They seem like passengers; some are even disenchanted... You sense that there is need to align the organizational processes with the business model.  The people with the journey. You know all the classic recipes - such as setting directions for your teams, establishing a strategic plan, defining performance indicators, putting processes in place, communicating with stakeholders... And yet, what happens when all of this has been done, and you still sense something is missing from the equation?  Are you being fascinating enough?

5. Influencing an audience – be fascinating!

Make the most effective use of preparation time by clearly identifying the objectives of your communications. Master public speaking tools. Transform a ‘hearing’ into a ‘listening’ audience – one which is mobilized and fascinated. 

Good communication does not come naturally to everyone.

Each of us uses words, expressions and attitudes which are unique to ourselves and which do not necessarily carry the same meaning for another person. Misunderstandings are therefore, in a sense, ‘normal’. We also make mistakes using media.  The wrong font, the slide that nobody can read.

Every message takes place on two levels: content & relationship

All communication conveys two kinds of information. One concerns facts, feelings and opinions: ‘content’. The other expresses something about the rapport between the people involved: ‘relationship’. This relationship often dominates the content.

The results of communication lie in the response we receive.

The reactions to our communication help us evaluate its appropriateness. If our primary expectations are not met, it is up to us, as communicators, to select a different mode of expression and a better way to fascinate the audience.

6. Building trust – how to do this …

Trust comes from delivering what you promised.  You can build trust by fascinating others with your past performance, track record of delivering results you commit to delivering, track record of your accomplishments.  Also, your experience in the business or a related one, your technical expertise, educational background, language, speech, patterns of speech, appearance, dress, grooming.  Don’t forget your personality, demeanor, and your professionalism demonstrated by your punctuality, organized approach, and manners.  Include as factors for fascination, the associations you have - company, contacts, other customers, clients.  Finally, the honesty, sincerity, and team player approach you exhibit.  Those are just a few factors involved in building trust.

7. Educating

A critical component in data science projects is the need to educate audiences in technologies that are increasingly complex.  This is not easy to do.  For instance, comparisons between SAS, R, Python, Perl.  Or discussion on how to turn unstructured data into structured data.  What is the best model a star schema or snow flaked schema?  What kind of statistical method to use logistic regression versus decision trees, or neural networks?

Often times, knowledge transfer takes the form of a continuous collaboration.  This helps the learning process because it is not a one-time shot.  Learning hardly ever happens in a single session! 

Understanding and transferring knowledge are complementary actions and are mutually enriching. Being a mentor could be a powerful launch pad for learning and personal development. 

Here is the link to the free fascination test. You’ll be surprised with yourself!

http://howtofascinate.com/get-my-profile Hope these ideas are useful to you!  Please like the post or better yet share it.  Let me know what your fascination profile is.  All the best.