Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesTo cross the analytics chasm, an organization must move beyond describing past performance and use data to predict likely outcomes and guide specific actions. The reliable route is not buying a bigger platform or running more proof-of-concepts. It is selecting a small number of economically important use cases, testing their feasibility, and delivering analysis at the detail and speed of the decisions it is meant to improve.
What the “analytics chasm” means
Bill Schmarzo uses the term to describe a capability shift: from retrospective business monitoring to predictive insight and prescriptive action. Reports and dashboards answer questions such as “What happened?” Predictive analytics estimates what is likely to happen; prescriptive analytics helps determine what should be done next.
The exact original page for the work titled Crossing the Big Data, Data Science and Analytics Chasm was not identified. The European Parliamentary Research Service cites a related Schmarzo article, “Crossing the big data analytics chasm,” dated September 25, 2018. That citation establishes a related publication, not that it is identical to the title here. The framework below is drawn from Schmarzo-attributed material, including his The Big Data Game Board™ article published by KDnuggets on November 19, 2018 and an author-attributed LinkedIn version.
The capability shift in practical terms
| Dimension | Retrospective monitoring | Predictive and prescriptive analytics |
|---|---|---|
| Question | What happened? | What is likely to happen, and what action should follow? |
| Data view | Aggregated summaries, often organized in conventional tables | Detailed histories that can be analyzed for individual people, products, services, transactions or devices |
| Data sources | Restricted internal, structured inputs | Broader internal and external sources, including structured and unstructured data where relevant |
| Timing | Batch reporting after an event or period | Timely analysis that can influence an operational decision |
| Output | Dashboard, report or scorecard | Prediction, recommendation, intervention or workflow change tied to a business outcome |
These are distinctions in Schmarzo’s framework, not a universal industry maturity scale. More data or faster processing is useful only when it improves a decision that matters.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →#1 Best Overall
Why organizations struggle to cross it
The value is economic, not merely technical
A data lake, model or streaming system is an input to a business result, not the result itself. The relevant question is whether an initiative can improve revenue, cost, customer experience, risk management or operational performance enough to justify its implementation and ongoing operation.
Technology experiments can outrun the decision
Schmarzo’s Big Data Game Board™ warns against allowing technology proofs-of-concept to carry exaggerated promises. A demonstration can show that a model or platform works in a narrow test; it does not prove that employees will use its output, that the necessary data is available at production quality, or that the resulting action creates value.
Rank #2
- Wiley
- Language: english
- Book - storytelling with data: a data visualization guide for business professionals
Business and data teams may optimize different goals
Business stakeholders understand the decision, constraints and economics. Data science and technology teams understand data quality, modeling, architecture and delivery. Crossing the chasm requires them to agree on the outcome and decision first, then design the analysis around it.
Scale creates focus problems
Organizations can identify more promising ideas than they can implement. Pursuing too many use cases at once spreads scarce subject-matter, engineering and change-management capacity across projects that may never reach operational use.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchPC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11A use-case-first route across the chasm
- Anchor the work in a material initiative. Start with a financial, customer or operational objective, such as reducing avoidable service cost, improving retention or increasing the reliability of an important process.
- Identify the decisions and drivers. Specify who makes the decision, when it is made, which factors influence it, and what action is available. This prevents a broad request for “more insight” from becoming an unfocused data project.
- Generate candidate use cases. Describe each candidate in terms of a decision, an expected outcome, the population or process affected, and the intervention the organization could take.
- Assess business value. Estimate the size and type of potential benefit—financial, customer or operational—without presenting an unvalidated estimate as a guaranteed return.
- Assess implementation feasibility. Check data availability and granularity, integration effort, model or analysis complexity, timing requirements, privacy and security constraints, ownership, and the ability to put an action into an existing workflow.
- Prioritize a manageable portfolio. Select a small number of high-value, feasible candidates rather than starting every attractive experiment. Make the trade-offs visible to both business and technical stakeholders.
- Assemble and analyze the relevant data. Collect only what the leading use cases require, at the level of detail and frequency needed for the decision. Validate definitions, history, quality and access before treating a result as operationally useful.
- Align the decision makers and delivery teams. Agree on the intended action, success measures, escalation path and accountable owner. A technically accurate output that nobody can or will act on has not crossed the chasm.
- Validate incrementally. Test whether the analysis is relevant to the business problem and feasible to operate. Use what is learned to refine, stop or expand the use case instead of assuming that a prototype is a finished solution.
How to prioritize competing initiatives
Use business value and implementation feasibility as the primary axes. A simple two-axis review makes the portfolio conversation explicit:
| Position | Meaning | Typical decision |
|---|---|---|
| High value / high feasibility | A credible opportunity with a realistic path to deployment | Advance first and define an owner, decision and validation plan |
| High value / low feasibility | Potentially important, but blocked by data, workflow, governance or delivery constraints | Fund targeted feasibility work; do not promise the full benefit yet |
| Low value / high feasibility | Easy to build but unlikely to materially change an outcome | Use selectively for learning, not as the centerpiece of the portfolio |
| Low value / low feasibility | Unattractive on both dimensions | Defer or reject |
The matrix is a decision aid, not a substitute for financial modeling or risk review. Value assumptions should be stated as assumptions and revisited as evidence arrives.
Rank #4
What “more granular and broader data” actually changes
Aggregate reporting can hide meaningful differences between customers, products, locations, devices or individual events. Detailed histories may reveal patterns that support individualized treatment or operational intervention. External, unstructured or otherwise broader sources may add context that internal tabular records lack.
That does not mean collecting more data automatically creates value. Each additional source introduces questions about permission, quality, lineage, integration cost, bias, security and whether anyone can act on the resulting signal. Data should therefore be selected against a defined use case and decision, not accumulated as an end in itself.
Free tools Windows power users keep installed
One-click scans. No signup required.
Best Value
Common failure modes and their corrections
- Starting with a platform purchase: define the business initiative and candidate decisions before selecting technology.
- Running an unconstrained experiment list: rank use cases by value and feasibility and limit the active portfolio.
- Promising a model’s result before testing deployment: separate a technical demonstration from validated business and operational feasibility.
- Using aggregate data for an individual decision: obtain the appropriate history and granularity, while checking privacy and governance requirements.
- Delivering a score without an action: identify the person, workflow or system that will respond and the conditions for intervention.
- Leaving ownership ambiguous: name accountable business and technical owners, along with the outcome and measures they share.
How to tell whether a project is crossing the chasm
A project is moving beyond reporting when its output is connected to a real decision and an available action. Review these questions at each stage:
- Is the targeted business, customer or operational outcome explicit?
- Can the team identify the decision maker, timing and intervention?
- Are the data’s history, granularity, quality and access adequate for that decision?
- Have business value and implementation feasibility been assessed separately?
- Is there an operational owner who can use the output?
- Has the result been validated for business relevance as well as technical performance?
- Are privacy, security, bias, reliability and implementation risks addressed before expansion?
Further reading
For a broader treatment of the economics behind this approach, Bill Schmarzo’s book The Economics of Data, Analytics, and Digital Transformation presents becoming value-driven and applying data and analytics economics use case by use case. It is related reading, not the exact work identified by the title of this article.
The practical test
An organization has crossed the chasm when analytics is no longer judged mainly by the existence of a dashboard, model or data platform. It is judged by whether a prioritized use case produces timely, trusted insight that changes a defined decision and contributes to a business outcome. That requires disciplined economics, workable data, and sustained collaboration between the people who own the decision and the teams that build the analysis.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
