Enterprise Data Strategy: Why Most Companies Get It Wrong

Enterprise data strategy is the set of decisions a company makes about how it collects, governs, integrates, and activates data to drive business outcomes. Done well, it connects marketing, sales, finance, and operations around a shared view of the customer and the business. Done badly, it produces dashboards nobody trusts, decisions nobody makes, and data teams nobody listens to.

Most large organisations sit closer to the second description than the first. Not because they lack data, but because they have too much of it in the wrong shape, owned by the wrong people, answering the wrong questions.

Why Enterprise Data Strategy Keeps Failing

I have sat in a lot of rooms where someone senior announces that the company is going to “become data-driven.” There is usually a slide. Sometimes there is a new VP of Data. Occasionally there is a very expensive platform purchase. And then, six to eighteen months later, the dashboards exist but the decisions have not changed.

The reason is almost always the same. The organisation treated data strategy as a technology problem when it was actually a commercial problem. They asked “what platform should we use?” before asking “what decisions are we trying to make better?” Those are very different starting points and they lead to very different outcomes.

When I was running an agency and we were scaling hard, we had a version of this problem internally. We had reporting. We had dashboards. We had data coming out of our ears. But the data was not connected to the decisions that actually mattered: which clients to invest in, which services were genuinely profitable, where our growth was coming from. It took a deliberate effort to rebuild the data architecture around the commercial questions, rather than around what was easy to measure. Once we did that, the data actually started influencing behaviour.

That experience shaped how I think about enterprise data strategy. It is not a data problem. It is a decision-making problem that data can help with, if you set it up correctly.

For a broader view of how data fits within commercial growth planning, the Go-To-Market and Growth Strategy hub covers the strategic frameworks that sit around it.

What Does a Real Enterprise Data Strategy Actually Include?

The term gets used loosely. Some people mean their CRM setup. Some mean their data warehouse. Some mean their analytics stack. A proper enterprise data strategy covers all of these, but it starts somewhere different: with a clear articulation of what the organisation is trying to achieve commercially, and which decisions sit between the current state and that outcome.

From there, a functional enterprise data strategy has five components that need to work together.

1. Data Governance and Ownership

Someone has to own the data. Not in a nominal sense, but in the sense that they are accountable for its quality, its definitions, and its accessibility. In most large organisations, this is where things fall apart. Marketing owns one version of the customer. Sales owns another. Finance has a third. None of them agree on what a “customer” even means, let alone how to count one.

Governance is the set of rules and accountabilities that prevents this fragmentation. It is not glamorous work. It does not make for a good conference presentation. But without it, every other investment in data infrastructure produces noise rather than signal.

2. Data Infrastructure and Integration

This is the layer most organisations focus on first, usually to their cost. Infrastructure matters, but it should follow the governance decisions, not precede them. The question is not “which data warehouse should we use?” but “what data needs to flow where, at what speed, for what purpose?” Once you can answer that, the infrastructure choice becomes much more straightforward.

Integration is the harder problem. Most enterprises have decades of legacy systems that were never designed to talk to each other. Customer data lives in the CRM. Behavioural data lives in the analytics platform. Transactional data lives in the ERP. Getting these to produce a coherent picture requires deliberate integration work, not just a new platform on top of the existing ones.

3. Data Quality and Definitions

Bad data is worse than no data, because it creates false confidence. I have seen marketing teams make significant budget decisions based on attribution data that was, on closer inspection, measuring almost nothing useful. The data looked clean. The reports looked authoritative. The decisions were wrong.

Data quality is not just about removing duplicates and fixing typos. It is about ensuring that what you are measuring actually corresponds to what you think you are measuring. This requires clear definitions, consistent collection, and regular audits. It also requires the organisational courage to say “we do not have reliable data on this” rather than producing a number that looks credible but is not.

4. Analytics and Activation

Analytics is where data becomes insight. Activation is where insight becomes action. The gap between these two is where most enterprise value gets lost. Organisations invest heavily in analytics capability, produce excellent analysis, and then watch it sit in a report that nobody reads or a presentation that influences nobody.

The activation problem is a change management problem as much as a data problem. It requires that the people making decisions actually trust the data, understand it, and have the authority to act on it. None of these are guaranteed, and none of them are solved by better visualisation tools.

5. Privacy, Compliance, and Ethics

This layer has become more important, not less, as regulatory frameworks have tightened and consumer expectations have shifted. A data strategy that ignores privacy is not just a legal risk. It is a commercial risk, because the trust damage from a data misuse incident can take years to recover from. Privacy should be designed into the data architecture, not bolted on afterwards.

The Attribution Problem That Nobody Wants to Talk About

One of the most consequential failures in enterprise data strategy is the way organisations handle marketing attribution. Attribution models are a perspective on reality. They are not reality itself. But in most large organisations, they get treated as ground truth, and the investment decisions that follow are built on foundations that are shakier than anyone admits.

I spent years working with performance marketing at scale, managing significant budgets across multiple channels. Earlier in my career, I was guilty of overvaluing lower-funnel performance metrics. The numbers looked good. The cost-per-acquisition looked efficient. The attribution model credited the right channels. And then I started asking harder questions about whether the performance we were measuring was actually being created by our marketing, or whether it was largely capturing demand that would have existed anyway.

It is a bit like a clothes shop: someone who walks in and tries something on is far more likely to buy than someone browsing the window. But the question is whether the shop created that intent or simply caught it at the right moment. Performance marketing often does the latter while claiming credit for the former. And attribution models, because they are built around measurable touchpoints rather than causal relationships, tend to reinforce this misattribution systematically.

The implication for enterprise data strategy is significant. If your data infrastructure is built around last-click or even multi-touch attribution, and if your investment decisions follow those attribution signals, you are probably systematically underinvesting in the channels that build demand and overinvesting in the channels that harvest it. The data looks clean. The logic looks sound. The strategy is wrong.

Forrester’s work on intelligent growth models touches on this tension between short-term measurement and long-term value creation. The organisations that get this right are the ones that hold their attribution data lightly and invest in a range of measurement approaches, including brand tracking, incrementality testing, and market mix modelling, rather than treating any single model as definitive.

Where Enterprise Data Strategies Go Wrong in Practice

Having been on both the agency and client side of data strategy projects, I have seen the failure modes repeat with remarkable consistency. They are worth naming clearly, because most of them are avoidable.

Collecting data without a use case

Organisations collect data because they can, not because they have a specific decision they need to make. This produces data lakes that are more accurately described as data swamps: vast, expensive, and largely impenetrable. The discipline of asking “what decision will this data improve?” before collecting it is rare but valuable.

Centralising without democratising

Many enterprises invest heavily in centralising their data infrastructure, which is sensible, and then fail to make that data accessible to the people who need to use it. The data sits in a warehouse that only the data team can access, and the commercial teams carry on making decisions based on spreadsheets and intuition. Centralisation without democratisation produces a data team that feels undervalued and a commercial team that distrusts data.

Confusing correlation with causation

This is the oldest problem in analytics and it has not gone away. In fact, as data volumes have grown and as machine learning has made it easier to find patterns in large datasets, the risk of acting on spurious correlations has increased. The discipline of asking “why?” before acting on a pattern in the data is essential, and it requires analytical rigour that goes beyond dashboard literacy.

Building for the current state

Data infrastructure built to support today’s business model often cannot support tomorrow’s. Enterprises that are growing, acquiring, or transforming need data architectures that are designed for flexibility, not just for the current state. BCG’s research on scaling agile organisations makes a related point about building systems that can adapt rather than optimise for a fixed state.

Treating data strategy as a one-time project

Data strategy is not a project with a start and an end date. It is an ongoing capability that needs to evolve as the business evolves, as the data landscape changes, and as new questions emerge. Organisations that treat it as a project tend to underinvest in the ongoing governance and maintenance work that keeps the infrastructure useful.

How to Build an Enterprise Data Strategy That Actually Works

The approach that works is not complicated, but it requires discipline in a few specific areas where most organisations take shortcuts.

Start with the commercial questions

Before touching infrastructure, governance, or tooling, spend time with the senior leadership team identifying the five to ten decisions that most influence commercial outcomes. These might be: which customer segments to prioritise, where to allocate marketing investment, which products to develop, which markets to enter. These decisions become the design brief for the data strategy. Every subsequent choice about what to collect, how to store it, and how to analyse it should trace back to one of these decisions.

This sounds obvious. It is not how most data strategies start. Most start with a technology evaluation or a data audit. Starting with commercial questions is the single change that most improves the odds of success.

Fix governance before fixing infrastructure

Agree on definitions, ownership, and accountability before spending on platforms. Who owns customer data? What is the canonical definition of a customer, a lead, a conversion? Who is accountable for data quality in each domain? These questions need answers before the infrastructure is designed, or the infrastructure will embed the existing confusion rather than resolve it.

Invest in the last mile

The last mile of data strategy is the point at which analysis becomes a decision. This is where most organisations underinvest. They build excellent analytical capability and then fail to ensure that the outputs reach the right people in a format they can act on, at the moment when the decision is being made. The last mile requires investment in communication, in training, and in the organisational processes that connect insight to action.

Tools like those covered in Semrush’s growth toolkit overview can support parts of this process, but the organisational design around how data flows into decisions matters more than the specific tools used.

Build measurement diversity into the strategy

No single measurement approach captures the full picture. A mature enterprise data strategy uses multiple lenses: attribution models for directional guidance, incrementality testing for causal inference, brand tracking for upper-funnel health, and market mix modelling for portfolio-level investment decisions. Each of these has limitations. Using them together produces a more honest and more useful picture than relying on any one of them alone.

Create feedback loops

The data strategy should include mechanisms for learning from outcomes, not just for reporting on activities. When a decision is made based on data, the organisation should track whether the predicted outcome occurred, and feed that learning back into the models and assumptions. Without this, the data strategy becomes a reporting function rather than a learning function. Hotjar’s work on growth loops illustrates how feedback mechanisms can be built into commercial processes, not just product development.

The Organisational Side of Data Strategy

The technology gets most of the attention in enterprise data strategy conversations. The organisational design gets almost none, which is backwards, because the organisational side is where most implementations succeed or fail.

There are three organisational questions that matter most.

First: where does data capability sit in the organisation? Centralised data teams produce consistency but can become disconnected from commercial reality. Embedded data capability within business units produces relevance but can fragment standards and governance. Most large organisations need a hybrid: central governance and infrastructure, with embedded analysts who understand the commercial context of their business unit.

Second: how does data capability connect to decision-making authority? Data teams that report into IT are rarely in the room when commercial decisions are made. Data capability that reports into the CFO or the CEO has a different status and a different relationship to the decisions that matter. The reporting line is not just an organisational detail. It signals what the data function is for.

Third: how does the organisation build data literacy across the commercial teams? The goal is not to turn marketers into data scientists. It is to ensure that the people making commercial decisions have enough data literacy to ask good questions, challenge weak analysis, and know when to trust a number and when to be sceptical. This requires investment in training and in the culture around how data is used in decision-making.

BCG’s thinking on commercial transformation makes a related point about the importance of capability building alongside structural change. You cannot separate the organisational design from the commercial outcomes you are trying to achieve.

Data Strategy and Go-To-Market: The Connection Most Teams Miss

Enterprise data strategy and go-to-market strategy are more closely connected than most organisations treat them. The decisions that a go-to-market strategy depends on, which segments to target, which messages to prioritise, which channels to invest in, which products to lead with, are exactly the decisions that a well-designed data strategy should be informing.

But the connection is often broken. The data strategy is owned by IT or a central data function. The go-to-market strategy is owned by marketing and sales. They operate on different timelines, with different vocabularies, and with limited mechanisms for the data function to influence the commercial decisions and for the commercial function to shape the data priorities.

Closing this gap is one of the highest-value things a CMO or CCO can do. It requires making data strategy a commercial conversation, not a technical one. It requires ensuring that the data function understands the commercial questions it is being asked to answer. And it requires that the commercial teams are willing to have their assumptions challenged by what the data shows, even when that is uncomfortable.

I remember being in a workshop early in my career where the founder handed me the whiteboard pen and left the room. The instinct was to play it safe, to summarise what had already been said. But the room needed someone to make a decision about direction, not to curate the existing thinking. Data strategy has the same dynamic. The value is not in producing more analysis. It is in making the call, based on the best available evidence, and being willing to be accountable for it.

There is more on the strategic frameworks that connect data, commercial planning, and growth in the Go-To-Market and Growth Strategy hub, which covers the broader architecture of how these elements fit together.

What Good Looks Like

Organisations with mature enterprise data strategies share a few characteristics that are worth naming, because they are different from what most data strategy frameworks describe.

They are sceptical of their own data. They know what their models can and cannot tell them. They hold attribution data lightly. They invest in multiple measurement approaches and they are honest about the limitations of each. They do not confuse confidence in the data with confidence in the decision.

They connect data to decisions, not to reports. The measure of success for their data function is not the number of dashboards produced or the volume of data processed. It is whether the decisions that matter to the business are better than they would have been without the data. This is a harder thing to measure, but it is the right question.

They treat data quality as a commercial priority, not a technical one. When data quality is poor, they escalate it as a commercial risk, not a technical debt. They invest in fixing it because they understand the cost of making decisions on bad data, not because a data team has flagged it as a hygiene issue.

And they build data literacy into their commercial teams. The goal is not a world where every decision goes through the data team. It is a world where the people making decisions are capable of engaging with data critically, asking the right questions, and knowing when the numbers are telling them something important. Forrester’s research on agile scaling points to data literacy as a consistent differentiator in organisations that successfully transform their commercial capabilities.

None of this requires the biggest data infrastructure or the most sophisticated tooling. It requires clarity about what the data is for, discipline in how it is governed, and the organisational will to connect insight to action. Those are leadership problems, not technology problems. And they are solvable, if you start in the right place.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions