There's a hidden cost that doesn't always find its way onto a budget, but every business analyst recognizes: the time it takes to ask a question, get a ticket, get a response, get a clarification, get a new estimate, and sometimes finally get data. I've started referring to it as the Business Analyst Tax - the cost that organizations pay (in time, delayed decisions, analyst dissatisfaction) whenever an analyst has to go through a human to get to an answer.
The scale of it is eye-opening when you see the research.
The Numbers Behind the Tax
Data analysts spend less than half their working day actually analyzing data. A global survey of around 500 data professionals found that over 60% had wasted time waiting for engineering resources multiple times per month, with many spending a third of every day struggling to gain access to data. The same study found that 68% of analysts had profit-driving ideas that went unimplemented due to a lack of time.
The prep and wrangling problem is just as bad: industry research broadly finds that analysts spend 60-80% of their time finding, cleaning, and organizing data before they can analyze anything. dbt Labs' 2024 State of Analytics Engineering survey found that over half of practitioners cited data organization as the most time consuming task in their week, citing poor data quality as their most common problem (57%) compared to 41% just two years previously. McKinsey estimates that knowledge workers lose roughly 19% of their time chasing and consolidating information.
These figures add up to a disturbing picture: the amount of time a business analyst spends on any given day is not spent actually doing analysis, but logistics.
What the Tax Actually Costs
The direct cost is obvious: take your analyst headcount, apply the time lost to waiting and wrangling, and multiply by fully-loaded salary. Most firms with a substantial analytics team have a meaningful number on an analyst time-loss metric like this at least.
The indirect cost is far more interesting, in my view: the decisions that didn't get made, or got made too late, because of long waiting periods for answers. "Can we break this out by region?" is a question that should take an afternoon. It becomes a week when you need that information to make a decision. The meeting it was needed for already happened; the decision got made, but without analysis.
There is also an opportunity cost that's rarely discussed: the work that doesn't get done, because an analyst is managing the queue. The Fivetran survey cited earlier found that 68% of respondents had ideas that would drive profit for their organizations but didn't have the time to pursue. This is not a fringe concern: it's about the most valuable work an analyst can do (interpretative, strategic work, pattern recognition) that should be happening, but doesn't, because the analyst is focused on fulfilling requests that should be automated.
Where the Tax Comes From
The Business Analyst Tax has two main sources, that compound each other.
The first is access friction: data is not in one place, and not accessible to the same people. Finding, wrangling, and understanding it requires institutional knowledge of where specific bits of information live and what it all means. When a business user wants to know something and can't find the data themselves, they have to open a ticket. The business analyst is the middleman between the question and the answer.
The second is context friction: even when a business analyst has access to the data, they have to spend significant time in order to understand what a question is asking and how to apply their knowledge to answer it. Without a shared understanding of how the concepts mentioned relate to actual data points and which ones matter, every data request comes with its own embedded research task.
Both of these are fundamentally context problems: the data exists; the question exists. The missing layer is the one that knows what the question means, where the answer lives, and how to get them together without a human in the middle.
What Changes When That Layer Exists
What's the before/after scenario that makes sense of this? An analyst who spends a third of their day waiting for access to data, and a third cleaning whatever they get, does not become more productive by working harder or hiring more people. They become more productive when access friction and context friction are gone: when a business user can ask a question and get an answer that's not just based on data, but on the right definition, from the right source, without having to file a ticket.
This is what happens when AI has organizational context: not just access to documents and tables, but a structured understanding of what things mean and what they don't, and which answer to what question is right. McKinsey estimates that AI-assisted knowledge retrieval could save 35% of the time currently spent on such tasks. Note that this is time saved just on knowledge retrieval - without taking into account the multiple steps that are eliminated when humans are no longer involved in the process.
In my view, the Business Analyst Tax represents one of the most immediate opportunities for enterprise AI right now: not because the technology is finally good enough, but because the problem is finally specific enough that an actual solution is possible. It's not a matter of making AI smarter; it's a matter of eliminating the round-trip between the question and the answer. It's an infrastructure problem, and it has an infrastructure solution.
The organizations that build the infrastructure that provides that missing layer of context will stop paying the tax. The ones that don't will continue to wonder why their analytics teams are always busy but never seem to stay ahead of the business.
