A standard best practice for any complex initiative that involves technology, change and business process redesign is to start small in order to achieve early success and visibility, which speeds adoption and enables the initiative to gain momentum.  This is why we separate business intelligence and data analytics engagements into phases and steps.

When we refer to a “phase” we mean the area of the association being analyzed – often this is the 5 stepshighest profile or highest revenue-producing areas of the association.  We recommend analyzing one business area at a time, such as membership, events, marketing or publications.  Once the focus is identified, we follow a standard set of steps.

After many years of performing successful business intelligence implementations for associations, we’d like to share our success checklist to help you understand the process:

Step One: Scope

  1. DSK’s CEO delivers an educational presentation, “What is Business Analytics and What Should I Expect?” This is generally delivered to a large audience within the client organization (often it is a full staff meeting).  The content for this presentation is developed in conjunction with the client.
  2. Kick off meeting(s) between the DSK team and client’s key stakeholders. The purpose is to ensure all participants understand the goals of the initiative.  Expectations and risks are discussed openly and the emphasis is on building rapport and open communication.  High level scheduling and responsibilities are discussed.  Often multiple meetings are needed to gain clarity and congruence. DSK generally assigns a team of two staff to our meetings – the Chief Analytics Officer guides the discussion and the business/data analyst captures the answers.  Sometimes a technical member of our team will attend key meetings in order to begin to assimilate information first-hand and begin logical data modeling activities.
  3. Document candidate business questions and prioritize them.  Delays in reaching agreement on the prioritization of business questions may occur and can stall deadlines.  Skillful negotiation is often necessary to help a team navigate through difficult business decisions and conflicting priorities, building consensus in the process.  We will help you!  DSK has a standard list of common association business questions grouped by area which can be used to frame the discussion.  Onsite discussions with your team will inevitably prompt additional questions regarding your organization’s objectives.
  4. Define deliverables and measurable outcomes for each specific business area (“phase”) to be analyzed.
  5. Create a timeline and high level project plan with milestones.
  6. Perform a high-level review of existing reports and queries to understand business focus, common data groupings and understand client’s business and data “language”

Step Two: Collect Data

During the collect step, DSK works with your technical and business staff to identify the relevant data sources to be included in the data initiative, and begins the data modeling and ETL processes for the data mart.  Steps include:

  1. Perform a technical architecture assessment.
  2. Perform a high level master data management assessment, which determines the authoritative source of record by business area and type. Note: master data management decisions can be political and difficult to make. It is essential that the team proceed expeditiously through this step or the engagement can get derailed.
  3. Create a data inventory, which documents data sources and meaning.
  4. Create a data dictionary, which documents the data types and meaning of the fields.
  5. Determine how much history to include.
  6. Create an inventory of existing, relevant reports (and the code that generates them) which is used to validate counts.
  7. Create a Common Language Dictionary to document the meaning of association-specific terms and metrics.
  8. Develop data mart to store the data to be used for analysis. This is iterative and will evolve.
  9. Configuring ETL (Extract, Transform and Load) process to populate data mart on a nightly basis with AMS/CRM data, as well as data from other sources, such as general ledger, Google Analytics, third-party event registration data, etc.  This is iterative and will evolve.

Step Three: Clean Data

During this step we profile data to identify anomalies.  It is very important at this time to identify the threshold level of ‘cleanliness’ that will be acceptable for the different types of data. Current data issues will be corrected and new processes may be created to proactively prevent future ones.  Steps include:

  1. Perform an assessment of the client’s data and develop a “report card” to quantify data quality issues.
  2. Perform data cleansing and de-duping based on data quality priorities.
  3. Document data that is missing or not available, but is needed for analysis.
  4. Identify sources/acquire external data to enhance understanding of the area being analyzed.
  5. Review data quality findings with relevant staff to determine the root cause(s) and identify which data to clean manually and which data can be cleansed through data quality scripts.

Step Four: Analyze Data

During the analysis phase DSK creates the visualizations that make the data come alive for business staff.  This is the fun part for most clients because this is the step where they can begin to “have a conversation” with the data – in a visual way!  Steps include:

  1. Design and create visualizations which answer the business questions established during the scoping phase.
  2. Deliver “The Analytical Mindset” training to staff, which enhances understanding of data analytics and speeds adoption by making it fun to see and understand data!
  3. Iterate through multiple versions of the visualizations to ensure appropriate level of detail is presented for each audience and that optimal flexibility for future analysis is inherent in the design.
  4. Soft-launch to internal core team

Step Five: Communicate Results

It is during this step that completed dashboards are presented and additional training provided to business and technical staff.  Feedback on the dashboards is collected and revisions made. Even at this stage the process is still iterative and we expect and even encourage any changes to the visualizations that will enhance the ability of staff to make better decisions using data.

Our experience has shown that a key predictor of a successful analytics initiative is a client that is fully engaged in the process and who has a solid understanding of how to make the most of their dashboards.  We integrate training into our process, carefully explaining the important features of Tableau and the functionality of each visualizations.  Candidates for training include the technical staff, business staff, analysts and those identified as Power Users, and any other individuals who will be using data to make decisions.

Steps include:

  1. Develop agenda and timeline for rollout and launch.
  2. Conduct onsite training to business staff and IT.
  3. Launch dashboards in conjunction with additional training.
  4. Review project goals and deliverables and celebrate success!
  5. Prioritize additional business questions which surfaced during the course of data discovery and plan subsequent phases

Continued regular on-site meetings with business users and IT staff after the initial analytics phase is essential.  We have learned that this is a necessary part of the journey for an association to become data guided.

Separating your business intelligence initiative into distinct business areas or “phases” and following the 5 steps above for each phase, is the best way we know to position your association for success with data analytics!