1. Establish Your Data Vision

We often hear leaders say: “We’d like to be more data driven,” or “We have a lot of data – it must be valuable!”. In our experience, these sentiments alone are too vague to form the foundation of a tangible data strategy. Invest time upfront to develop a clear vision for your data program.

For growth stage companies, data strategies are often oriented around one or more of the following objectives:

• ‍Operational excellence: Do your data provide insights that might improve staffing or inform scheduling decisions? Is there information that is relevant to your supply chain logistics? Or, could add efficiencies to your product delivery process? Data can be a powerful tool in managing costs, improving gross margins and driving operational excellence.
  ‍
• ‍Revenue growth: Do your data offer information about customer characteristics or segmentation? Many successful data programs support new customer acquisition, conversion, retention, and growth of existing markets. Data can help you identify your most profitable customer acquisition channels. It can highlight the characteristics of prospects that are most likely to convert and identify those that may remain over the long-term. This information can, in turn, inform plans to target new markets and/or customer segments.
  ‍
• ‍Customer insights: Sharing data with your customers – in the form of aggregated usage charts, benchmarks or raw data provided via API – can help make your value prop more clear and, ultimately, strengthen your customer relationships. This is particularly true for B2B companies – though we have seen several consumer businesses find creative ways to engage customers with data as well.
  ‍
• ‍Algorithmic products: Data can support dynamic, personalized product experiences. Recommendations (suggestions about which movie to watch next, which accessories match an outfit, etc.), price predictions, credit or fraud scores, and personalized discounts are all widely known examples of machine learning-based algorithmic products.
  ‍‍
• Data monetization: Do you have unique data that other businesses value? For some businesses, the data itself is the product, while in others, data can become an additional product line offered to new types of customers. Consider whether there is unique data generated in the regular course your business, or whether there is data to be captured through proprietary methods. Successful data monetization requires both a customer set that will value the data and the underlying technology and processes to reliably deliver high-quality data feeds in a timely manner.

It is important to note that these categories refer to “offensive” (revenue generation or profitability improvement) rather than “defensive” (harm-reduction) objectives, reflecting a shift in understanding the role data plays in organizations today. We encourage companies to consider each of the above as you craft your data strategy. If more than one resonates with your company’s goals, rank order them based on potential impact on your business. Remember: the first component of a well-crafted data strategy is a clear vision; this approach may help bring that vision into focus.

2. Assess Your Current State

You’re likely unsatisfied with the current state of your organization’s data; that may be why you’re creating a more concrete data strategy in the first place. With a vision established, you can turn your focus to understanding the maturity and accessibility of your data.

When we think about data, many operators think first about technology (i.e., the source of the data), but we recommend an honest, fulsome assessment – a structured inventory of the people, processes and technology across each of your objective areas – to support a more concrete action plan.

• ‍People: Ultimately your data strategy will be carried out (or not) by people in the organization. Take stock of your data team – including data analysts, data scientists and data engineers – in terms of numbers, skillsets and relevant experience. Understand, too, the data literacy of the rest of your organization. Are teams across your organization data literate? Do they view data as central to success? If your data vision includes the possibility of externally facing data products, your assessment should include customers or partners as well. How data savvy are your customers today? Is your team set up to build the right product(s) and provide the right training to those customers?
  ‍‍
• Process: How are decisions currently made within your organization – from the individual contributor to the boardroom – and what role does data play (or not play)? In what situations are people skeptical of, or threatened by, data? Is there a product development process that can incorporate new data products? What business processes might change, and who has a say? Developing a clear picture of your data culture – including the product development and business processes needed to support new data products or use cases – is often the biggest factor in the success of any data strategy. (Read more about a culture of experimentation here)
  ‍‍
• Technology: With a more complete understanding of your people and processes, you can turn to technology (i.e., the source of your data). Does your data live in spreadsheets, internal databases, SaaS tools, streaming feeds, etc.? Are data sources integrated into a data warehouse? What, if any, ETL (extract-transform-load), data warehouse platform, or BI / reporting tools are in place? Your data infrastructure may include machine learning platforms, data catalogs observability tools, reverse ETL, or a variety of other related technologies as well. It’s essential to assess your current data sources and technology infrastructure, including advantages and shortcomings.

Assessments often start – and end – with frustrations over data quality. While this is certainly critical, building a more structured understanding of the current state may help develop a strategy that can unlock (and help maximize) the true value of your data.

3. Develop Your Roadmap

With a clear vision and honest assessment of your organization’s data in hand, you can begin crafting a roadmap, including a path to some “quick wins” for short term impact and some foundational investments to support longer-term success. A roadmap can (and should) remain high-level – save the tactics for another document – but often includes a combination of the following elements:

• ‍A plan for your people: A thoughtful org structure, and the appropriate resources to lead the effort are critical. You may need to create a new team, or change reporting structures to better align resources with your data objectives. If you don’t already have one, consider hiring a data leader. New objectives may require skills and experience that don’t already exist in the company. Customer-facing data products usually benefit from dedicated product management as well. Training and enablement should be included in the plan. Employees beyond the data team – including individual contributors, executive leadership and board members – need to understand and interpret data.
  ‍‍
• A plan for your processes: Manual processes may need to become automated; software that encodes processes may need to be updated. Responsibility for these often lives outside of the data team, so developing partnerships and establishing the common benefits are key. Ways of working usually need to adjust as well. Teams or functions that have not previously worked together may need to update their expectations of whose work they depend on, and who depends on them. These changes are often overlooked but can make or break the data strategy.
  ‍‍
• A plan for your data infrastructure: The data warehouse, ETL pipelines, and BI tools may need to be updated (or set up for the first time). You may favor best-of-breed or open source tools, or look first to ecosystems already familiar to your IT department or engineering team. Infrastructure to support experimentation should also be part of your stack. Data quality improvements should have a place in your plan. Data needs to be “fit for purpose” and improvements such as resolving customer identities across data sources, refining definitions of active and churned customers, or increasing refresh frequency are often prerequisites to further work. Data productivity tools that improve data quality and observability, catalog data, provide automation for machine learning, or enrich operational systems can increase employee job satisfaction and the value of your data assets at the same time.