Data-Driven Consulting: Why Advice Without Numbers Is Just an Opinion
Owners of small and mid-sized businesses have accumulated a well-earned distrust of consulting. The standard scenario is familiar: a consultant arrives, runs interviews, delivers a slide deck of recommendations a month later, collects the fee and disappears. The deck goes into a folder. Six months later no one can say whether the recommendations worked — because no one measured. Here is how AI analytics rebuilds every link in that construction.
Why the classic genre does not work
The genre's problem is not that consultants are stupid. The problem is that the entire cycle is built on opinions: the diagnosis is made from conversations, the recommendations are not tied to measurable quantities, and verification of results is not part of the service's design at all.
Each of the three links — diagnosis, recommendation, verification — can be rebuilt on data. That is what separates consulting worth paying for from the genre of beautiful presentations.
Diagnosis: from data, not interviews
The first thing that changes is the source of the diagnosis. Instead of tell us how your sales are organised — a connection to the company's actual data: the till, the CRM, the ad accounts, maps and aggregators. Modern tools allow such a diagnostic picture to be deployed in days rather than months, and it regularly diverges from the team's account — not because the team lies, but because people see their own average while the data shows the distribution.
The typical findings of such diagnostics recur from company to company in our audits. A channel considered profitable in fact breaks even once every cost is honestly counted. A location or line of business dragging down the aggregate while staying invisible in averaged reports. Hours and days when inquiries are systematically lost — evenings, weekends, peak load.
Business diagnostics: the interview approach versus the data approach
Indicative timelines for a small or mid-sized business. The difference is not only speed: a data-based diagnosis is verifiable and does not depend on what the team said about itself.
Recommendations: tied to a metric
The second change is the form of the recommendations. In the data paradigm, a recommendation does not sound like strengthen your work with the customer base. It sounds like: here is the metric, here is its current corridor, here is the target, here is the action expected to move it. The difference is fundamental: such a recommendation can be verified, which means it can be answered for.
This disciplines the consultant too. When it is known that the result will be visible in a dashboard, everything unfalsifiable disappears from the recommendations — every improve loyalty and work on awareness that can be neither executed nor failed.
Verification: consulting with an open ledger
The third — and main — change: consulting acquires an afterword. If the company runs an owner dashboard, the effect of every implemented recommendation is visible in the same numbers the diagnosis was made from. Advice stops being a one-off transaction of money for a slide deck and becomes a cycle: diagnosis, action, measurement, next diagnosis.
For the owner this inverts the economics of the service. Paying for an opinion is a risk. Paying for a cycle in which every link is verifiable is an investment with visible returns. To see the instrumental basis of such a cycle, open the owner dashboard demo at mrktr.pro/owner-intelligence — the same screen in which both the diagnosis and the verification live.
What to ask a consultant before signing the contract
Four questions that separate data-driven consulting from the genre of beautiful presentations.
What data will the diagnosis be made from — interviews, or a connection to the company's actual systems? In what form will the recommendations come — bullet points, or metric-action-expected-shift triplets? How will the result be measured — and will the company keep the measurement instrument after the project ends? And what does the consultant propose not to do — the ability to strike things out is what distinguishes prioritisation from a wish list.
The last point deserves a separate comment. The main value of an outside view for a small company is not new ideas but the order of priorities: out of twenty possible improvements, resources will cover three, and the cost of choosing the wrong three is higher than it appears. This is exactly where data is most useful — it turns an argument about priorities into a calculation.
Key Takeaways
- 01Classic consulting is built on opinions: diagnosis from conversations, recommendations without metrics, verification absent from the design.
- 02Data diagnostics takes 5–10 days against 4–6 weeks for an interview audit — and its conclusions are verifiable.
- 03The working form of a recommendation: metric, current corridor, target, action. Everything unfalsifiable filters itself out.
- 04A dashboard turns consulting from a transaction into a cycle: diagnosis, action, measurement, next diagnosis.
- 05The main value of an outside view is not ideas but the order of priorities; data turns an argument about priorities into a calculation.
- 06Four questions before the contract: the source of the diagnosis, the form of the recommendations, the method of measurement, and what is proposed NOT to be done.