DECHY logo

Ctrl K or / to search, arrows to move, Enter to open

Solution

Business dashboards

A good dashboard answers a manager's morning questions without anyone building a report. We design dashboards around the decisions people make, connect them to your own systems, and settle what each number means so everyone trusts it.

  • 1,200+ clients helped
  • Building since 2010
  • 15 services, one team
  • 50+ technologies
A candlestick chart on a dark screen

At a glance

What we build

We build business dashboards: one screen per role showing the few numbers it needs to run its part of the business, taken straight from your own systems and always current.

The problem

Numbers that take a week to get and an hour to argue about

On Friday someone exports sales from one tool, costs from another, and stock from a third. They paste them into a spreadsheet, fix the dates, and send it round. By Monday it is out of date. Two managers quote different revenue figures, because each counted a little differently.

That is not a technology shortage. It is a definitions shortage. Until the business has agreed what counts as a sale, an active customer, or a late order, no dashboard can be trusted, however good it looks.

So the work has two halves. One is technical: connecting the sources and keeping the data current. The other is the harder one: agreeing the measures. We do both, and we start with the second.

How we solve it

From every figure to the few that matter

Stages that narrow: Everything you could measure: Hundreds of figures across many systems.; The decisions people make: Each role asks a handful of recurring questions.; Agreed measures: One written definition for each number that matters.; A dashboard for each role: Few numbers, clearly shown, always current..

  1. Everything you could measureHundreds of figures across many systems.
  2. The decisions people makeEach role asks a handful of recurring questions.
  3. Agreed measuresOne written definition for each number that matters.
  4. A dashboard for each roleFew numbers, clearly shown, always current.
This diagram shows four stages that narrow, from everything you could measure through the decisions people make and the agreed measures to a dashboard for each role.

What it looks like

What a dashboard for one role looks like

A dashboard for one role leads with a few tiles, one trend, and the items that need attention.

A data dashboard on a screen

Routes

Three ways to build a dashboard

  • A reporting tool on your data

    Suits

    Internal views for staff, quick to build, with standard charts.

    Watch for
    Per-user licensing, and limits when you need unusual layouts or sharing outside the business.
  • A custom-built dashboard

    Suits

    Role-specific screens, customer-facing views, and dashboards inside a product.

    Watch for
    More to build, but full control of design, access, and cost at scale.
  • A spreadsheet connected to live data

    Suits

    A first version or a small team with a few sources.

    Watch for
    Hard to secure and to share, and fragile as formulas multiply.

Worth knowing

What makes a number trustworthy

A number is trustworthy when anyone can say where it came from, what it counts, and who owns it. A dashboard cannot supply that. The work around it does.

  • One written definition for each measure
  • A named owner who can explain a change
  • A check against a figure the business already knows
  • A visible time of the last update

What to build

Which kind of dashboard do you need?

Dashboards differ by who uses them and how often. Four kinds cover most requests.

Who is the dashboard for?

Choose this when

Owners and executives who want revenue, margin, cash, and pipeline on one page.

What you get
A short, stable set of measures with clear trends and comparisons.
Keep in mind
Keep it to a few numbers. Every addition dilutes the rest.

What you get

What a dashboard project delivers

Trusted numbers

  • A written definition and owner for every key measure.
  • Checks that the dashboard matches your known figures.

Current data

  • Automatic refresh from your own systems.
  • Alerts when a number moves outside its normal range.

Clear screens

  • A dashboard for each role, designed for quick reading.
  • Views that work on a phone as well as a desk.

Safe access

  • Each person sees only what their role allows.
  • Logs of who looked at what.

How we do it

From questions to a dashboard people use

  1. Start from decisions

    We ask each role which decisions they make and which numbers they would need to make them faster.

    What it produces

    • Decision list
    • Draft measures
  2. Agree the definitions

    Each measure gets one written definition, an owner, and a source. Disagreements are settled here, not in the dashboard.

    What it produces

    • Measure definitions
  3. Connect and clean the data

    We connect the sources, check the data against known figures, and fix gaps. The data layer is built to refresh on its own.

    What it produces

    • Data connections
    • Reconciliation report
  4. Build, test with users, and release

    We build the first dashboard, review it with the people who will use it, adjust it, and release it with training.

    What it produces

    • Dashboard
    • User feedback
    • Training

Common questions

Questions about business dashboards

How many numbers should a dashboard show?

Few. Most roles need between five and ten measures to run their part of the business. More than that and the important ones are lost.

Can a dashboard use data from several systems?

Yes, and that is the usual case. The data is brought together in one place, matched, and checked before it is shown.

Do we need a data warehouse?

Not always. A modest set of sources can sit in a well-designed database. A warehouse is worth it when volumes or the number of sources grow.

Why do you insist on agreeing definitions first?

Because a dashboard shows numbers faster. If two teams count differently, it only makes the disagreement more visible.

Can we keep our existing reporting tool?

Often, yes. We can build the data layer underneath it, so it shows trusted numbers, or recommend a change if it is the limit.

Have a different question?

Tell us what you are weighing up and we will answer it plainly.

Ask a question

Already a client? Email support@dechytech.com

Technical depth

How we build the data behind it

The detail a technical reviewer will ask about: the data layer, the dashboard itself, and access and audit.

  • Source data is copied on a schedule into a reporting store, so dashboards never slow the systems that run the business.
  • Raw data is kept as received. Cleaned tables sit on top, and each measure is calculated in one place, as tested, version-controlled SQL.
  • For larger volumes we use a warehouse such as BigQuery; for smaller ones, a well-designed PostgreSQL database is enough.

Tell us which decisions are waiting on numbers

Describe the decisions and where the figures come from today. We will outline the measures, the data, and a first dashboard.