Build and implement a data strategy

1. Write down your company strategies, goals and determine the main use-cases you intend to use data to achieve those goals

In a spreadsheet, start with the main strategies as columns. Under each column, list potential goals that play into that particular strategy. Then, list the ways below the goal in which data can help you achieve those goals (in any way). For example, for the strategy “achieve higher saturation in existing markets”, for the goal “achieve more B2B sales” you could list “automate lead scoring” and “gather qualitative data from prospects” to the possible data use-cases.

2. Interview key stakeholders such as Marketing Analysts, Data Engineers, Sales Reps to determine their pain points when applying data to their day-to-day business problems

Force stakeholders to assess which pain points are most critical and which are less important. For example, “reports not in real-time” is not as important as “subscription reporting not 100% accurate.”

3. Match the stakeholder pain points from step 2 with your data strategic goals from step 1 and assess the gaps

Take both lists from step 1 and 2 and compare the gaps – is there a clear overlap in how you see the strategic goals for data and how stakeholders struggle to apply data to their business problems? No – the goals you listed are not aligned with the actual problems the business is facing! Pause here and invite your stakeholders to speak with you again to ensure alignment and then you can move on to the next step. Yes – then you’re ready to move to your strategy for data collection by building a measurement plan.

4. Write a measurement plan by starting with an assessment of what data sources are needed for each use-case

In your original spreadsheet, write down what kind of data is needed for each use-case you defined in the goals. For our previous example “achieve more B2B sales” for the use-case “automate lead scoring” you could list “salesforce”, “website behaviour” and “third-party company metadata” as data sources. List all possible sources at this point.

5. Choose the appropriate data collection & processing methodology together with data/analytics engineering

Choose which data processing methods (e.g. custom APIs, ETL/ELT) and what architecture of data warehousing (e.g. Redshift, BigQuery) you want to use together with your data/analytics engineering resources. If you’re going to be collecting behavioral data from your product, apps or other sources (like Smart TVs) you will definitely need to plan for a proper implementation of a data collection.

6. Rewrite and define your data outcomes based on the use-cases you set in step 1

As everything is in place for data collection and processing, the data outcomes can now be formulated concretely. For example, “automate lead scoring” could be rewritten and defined as: “Unification of Salesforce account & contact information, behavioral website data and external third-party data in order to automate and improve the scoring of a lead’s potential to convert to a paying customer based on a scale of 0-100” Each data outcome will need to be individually planned and addressed in your organisation.