Dashboards & BI

Dashboards that say
what changed

We design and build dashboards around the decisions your teams make, with each number set against its usual range and said in plain words, so a glance shows what changed, and whose it is.

Thanks, we’ll be in touch soon.

{{ heroStatus }}

The drawing: one dashboard with three charts, Orders, Returns and Late deliveries. Under each chart a sentence says what it shows: two read “As usual.” and the third reads “Above its usual range.” That chart is then shown as a table of the same numbers.

{{ heroSay }}

You’re in good company

A few of the companies we’ve worked with.

{{ wallSay }}

Plenty of dashboards, few decisions

A dashboard, a screen of numbers and charts that updates on its own, is easy to build and hard to decide from. Over time a screen fills with every number someone once asked for, each with an arrow for whether it rose or fell since last week. Numbers rise and fall on their own, so an arrow alone doesn’t say whether anything changed. Nothing on the screen says which number matters today, or whose it is. So people stop opening it, and ask someone for a spreadsheet instead.

A dashboard, in an example, drawn as a grid of tiles. Every tile has an arrow, some up and some down, each compared with last week. Nothing marks which one matters, or whose it is.

What we build

The screen, and what makes it worth opening.

  • One screen for each decision

    Each dashboard starts from a decision someone makes, how often they make it and what they do next, written down before anything is built. What doesn’t help that decision comes off the screen.

  • A comparison that fits

    Each number set against the same time a year before, a target or its own usual range (how far it ordinarily moves), chosen for what the number is, not left at last week by default.

  • Changes that are news, marked

    Each number’s usual range worked out from its own history, so an ordinary week looks ordinary and an unusual one stands out.

  • An owner and a next step

    Each number that matters (often called a KPI) has a named owner on your side, and a written next step for when it leaves its range. No number goes on a dashboard without them.

  • Said in words

    Each chart carries a plain sentence saying what it shows, so nobody has to work it out in the meeting.

  • Read without relying on color

    Each status is a word and a shape before it’s a color, and the lines and bars you need stand out clearly from what’s next to them.

  • The table behind every chart

    Each chart’s numbers as a table that a screen reader, a search or a spreadsheet can read, and an export that carries the date of its data.

  • The monthly report, off the spreadsheet

    A report is the same numbers, sent or printed for one date. The one someone builds by hand is built once and refreshed on its own, with the same numbers as the dashboards.

  • Fewer, and used

    Use measured, so the dashboards people rely on stay current and the ones nobody opens are retired once their owner agrees.

Pipelines, a warehouse and agreed definitions? Data Engineering →

Dashboards for your customers, inside your product? SaaS Development →

How your software behaves for the people using it? Observability & SRE →

A change, or the usual movement

A dashboard that marks every rise and fall teaches people to react to noise, or to stop looking. So we show each number against its own usual range. An ordinary week looks ordinary. An unusual one stands out, and goes to its owner.

An exampleLate deliveries were above their usual range last week.

Last week: Outside its usual range. Worth a look.

Usual range

Earlier weeksLast week

The same chart, as a table
WeekWhere it sat
The earlier weeksInside the usual range
Last weekAbove the usual range

Owner: Operations. Next: ask the carrier.

Every number moves from week to week, even when nothing has changed. A range worked out from its own history shows how much movement is ordinary.

A week outside the range is a reason to look for a cause, not proof of one. Where the range sits is a choice: narrower finds more, and raises more false alarms.

A long run of weeks on one side of the middle is worth a look too, even inside the range.

Where your business has seasons, a number is compared with the same time a year before.

Read without relying on color

Dashboards lean on red, amber and green. Some people can’t tell those colors apart, and a projector or a printout can lose them. The web’s accessibility standard (WCAG) asks that color is never the only way information is shown, and that the parts of a chart you need in order to read it stand out clearly from what’s next to them. So each status is a word and a shape first. Lines and bars are named where they sit, not in a key to one side. And every chart has its table behind it, for a screen reader, a search or a spreadsheet.

For an audit of a whole site or product against that standard: Accessibility →

Three stores and the status of each, in an example. Each status is a word with its own shape.

Stores this weekAn example

  • Market Square: On track
  • Hill Road: Watch
  • Ferry Lane: Needs a look

Each status is a word and a shape. Printed in gray, it reads the same.

Our work

Every family the office looks after has a different idea of what reporting means.

Lysander Private Office

A private reporting portal that gives each family its position and allocation, its mandate in plain language, its statements and signatures in one place, with its partner a tap away.

See it on Our Work →

Start with one screen

We begin with the one screen your team looks at most, or the report someone builds by hand each month. We ask what’s decided from it, by whom and how often, and which numbers on it nobody uses. Then we rebuild that one screen with you: fewer numbers, each against its usual range, each with an owner, each chart with its sentence and its table. You get that screen in use, a short written note on how each choice was made, and a list of the screens to do next, in the order we’d do them.

Your dashboards and the data behind them stay in your accounts and tools. Each person sees only what their role allows, and personal data is left out where the question doesn’t need it. Where our engineers reach the data, it’s for a task you’ve agreed, and it’s recorded.

Work with us as a dedicated team, on a fixed price project or a mix of both. Compare engagement models →

How each is paid for: Pricing →

One screen and the label under it, in an example. The label says what the screen decides, when it’s read and who owns it.
  • Decides: what to reorder
  • Read: every Monday
  • Owner: Operations

Questions worth asking

Which BI tool should we use?

BI is short for business intelligence. Use the tool you already have, where it does the job: a new tool rarely fixes a dashboard nobody reads. If you’re choosing one, we compare them against your own questions: who reads, how often, on what screen, and what it costs you to run. We don’t sell or resell any of them, and we take no commission from one.

Do we need a data warehouse first?

Not always. For a few sources, reports can read from a copy of your database. When several systems have to agree, the data is brought together first: Data Engineering →

Where do agreed definitions come in?

First. A dashboard can only be as right as the numbers under it, so what each number means is agreed and written down before it goes on a screen. That work is described in Data Engineering →

How do you work out a number’s usual range?

From its own history. Where there isn’t enough history yet, the number is shown without a range, and says so. A number that follows the seasons, or climbs steadily, gets no range either: it’s compared with the same time a year before. The range is yours to set: a narrower one finds more, and raises more false alarms. A week outside it is a reason to look, not proof that something is wrong.

Can people still export to a spreadsheet?

Yes. People will always want to work with the numbers themselves. Every chart has its table, and an export carries the date of its data, so a spreadsheet starts from the same numbers as the dashboard.

Can our customers see dashboards inside our product?

Yes. Each customer sees only its own data, checked on every path the data can leave by. SaaS Development →

Are the dashboards accessible?

Where we build the charts ourselves, we build them to the current version of WCAG, at level AA. Inside a BI tool, we use what the tool offers, tell you where it falls short, and give the table alongside. Accessibility →

Who looks after them afterward?

Either. We can look after them with you: keep each one current, add numbers as your questions change, and retire what nobody opens once its owner agrees. Or we hand them over to your team, with the written note on how each choice was made. Support & Maintenance →

What does it cost?

It depends on how many screens you have and where their numbers come from. The first call is free. Your proposal sets out the scope, the price or rate, and how you pay, before any work starts. We promise nothing about revenue, growth or savings. Pricing →

Let’s connect it better

Tell us which screen your team looks at most. We’ll suggest where to start.

Thanks, we’ll be in touch soon.

{{ ctaStatus }}