Data visualization consulting
Data visualization consulting for dashboards people actually open
The dashboards got built. There was a launch, and a training session. Six weeks later everyone is back to asking an analyst, or quietly maintaining their own spreadsheet version of the same number.
That is rarely a charting problem, which is why buying a better BI tool usually does not fix it.
5 years
Building analytics that runs in production
10+ years
Minimum experience, every consultant on your project
What data visualization consulting buys you
You are hiring a team to design and build the reporting your business makes decisions from, instead of buying a BI tool and working it out between other priorities. That covers agreeing which decisions the reporting is for, modelling the data so each metric means one thing, and building the dashboards in your own tool.
The charting is the last and fastest part. What takes the time is deciding what to leave out, and getting two departments to accept a single definition of a number they have each been reporting differently for years. Skip that and you get a beautiful dashboard nobody believes.
Judge any proposal on what happens after handover. If your team can change a metric definition without calling anyone, it worked.
When to hire, and when not to
Working this out now costs you an hour. Working it out after the build costs you the budget and another year of people not trusting the reporting.
Worth hiring for
- You have dashboards and nobody opens them. The build happened, the launch happened, and people went back to asking an analyst.
- Every leader has their own spreadsheet version of the same number, and meetings start by arguing about which one is right.
- The reporting exists but takes days to change, so by the time a question is answered the decision has moved on.
- You are about to buy a BI tool, or you bought one last year and it has not paid for itself yet.
Probably not yet
None of these are things we cannot build. They are situations where a dashboard is the wrong shape for your stage, and it would not earn back what it costs you.
- –Your processes are still changing month to month. A dashboard encodes assumptions about how you operate, so when those keep moving you spend more time maintaining it than reading it.
- –There is nobody who would own it. Reporting needs a person who notices when a number looks wrong, and on a small team that job tends to belong to nobody in particular.
- –One person needs one answer once. That is a query, not a dashboard, and dashboards built for a single question get abandoned.
- –You are pre-seed, or pre-operations. With no ops function and few settled processes, the data is usually financial only, and your accounting tools already report that well.
What data visualization consulting services cover
Data visualization consulting services get quoted as a dashboard count almost everywhere, which is the wrong unit. Your scope would be drawn from some of this and never all of it, and the first item is the one most often missing from a quote and the one that decides whether the rest works.
- Working out which decisions the reporting is meant to support, before anything gets designed. This is most of the value and the part usually skipped.
- A semantic or metric layer, so revenue means the same thing in every chart and nobody has to reconcile two numbers by hand.
- Dashboards built for the people who will actually open them, which usually means fewer charts than you expect and a clear default view.
- Self-serve setup, so the common follow-up questions get answered without a ticket.
- Performance work, because a dashboard that takes thirty seconds to load is a dashboard nobody uses twice.
- Rebuilds of reporting that already exists but has stopped being trusted.
- Documentation and handover as deliverables with a date, so your team can change it without calling us.
Working in the BI tool you already pay for
If you are mid-way through a tool evaluation, the honest news is that this decision matters less than the vendor demos imply.
The tool matters less than the layer under it
Power BI, Tableau, Looker, and Omni all produce good reporting when the model underneath is right, and all produce arguments when it is not. Most reporting problems people blame on a BI tool are modelling problems.
Looker and the semantic layer
Looker consulting work is usually LookML: getting the model, the explores, and the permissions into a state where self-serve is safe rather than merely enabled. The same principle applies in Omni and in Power BI's semantic models.
Built in what you already own
If you are already paying for a BI licence, that is almost always the right tool to build in. Switching costs more than it saves unless the current one genuinely cannot do the job, which is rarer than vendors suggest.
The warehouse underneath matters more than any of it. If the reporting is slow or the numbers disagree, that is usually where the problem lives, and data warehouse consulting is the piece that fixes it.
Who your data visualization consultants would be
Most firms sell you one team and staff you another. You will meet someone senior during the sale, so ask whoever you are talking to, including us, whether that person is still on the project in week six.
- A US-based senior data leader designs the engagement and stays your point of contact throughout. You meet that person before you spend anything.
- The same engineers work in your BI tool and your warehouse week to week, so you explain your business once rather than to a new face every month.
- Every consultant on the project has ten years of experience or more, which means none of your budget pays for someone to learn dashboard design on your data.
- If your team wants to build part of it themselves, that is a normal way for this to go and it lowers the price.
What to ask any data visualization consultant before you sign
Put them to us as readily as to anyone else on your list. The answers separate firms faster than the proposals do.
- Which decisions is this reporting for? A data visualization consultant who starts with chart types rather than decisions has skipped the part that matters.
- What will you tell us not to build? Every useful engagement removes charts as well as adding them.
- How will we know if people actually use it? Agree on that measure before the build, not after.
- Who owns the semantic layer when you leave, and can our team change a metric definition without you?
- What happens if the data turns out to be wrong mid-project? Find out who absorbs that before you sign.
Questions about data visualization consulting
The things people ask before the first call.
Data visualization consulting is hiring a team to design and build the reporting your business makes decisions from, rather than buying a BI tool and working it out yourself. In practice that means agreeing which decisions the reporting supports, modelling the data so each metric means one thing, and building dashboards in your own BI tool that people open without being told to. Most data visualization consultants sell it alongside the modelling underneath, because a chart on unmodelled data is where these projects usually fail. The engagement should end with your team able to change it.
Get a scope and a price before you commit anything
A short working session with the senior lead who would run the engagement, and one specific problem priced in writing. Yours to act on however you like, including elsewhere.
