👋 Hey, it’s Sundar! Welcome to experiMENTAL: a weekly newsletter on B2C marketing & data science how-to guides, frameworks, and stories from 15 years including early Uber.

Death by 1,000 dashboards.

Every week, I see a few posts on my LinkedIn with the same sentiment about dashboards as the screenshot above. My career experience resonates with this 100%. Over time, I’ve learned that # of dashboards is inversely correlated to impact of dashboards.

For those of us in data science & analytics, we’re tired of building dashboards that end up in the grave yard. On the flip side, stakeholders are tired of being disappointed by dashboards that don’t meet their expectations.

So, below, I’m going to share the strategy you should use to make dashboards that are impactful and don’t collect dust after a week.

Note: This article will not talk about the technical components of a dashboard. It won’t cover tools, data sources, pipelines, etls, etc. If you came here for that, I’m sorry.

However, if you came here because you’re a founder or Marketer looking to create a better data driven culture in your company by creating better dashboards then keep reading.

But, first, a fun fact.

The origin of the term dashboard

The word “dashboard” was originally used to describe the wooden board attached to the front of carriages to prevent mud and rocks from being splashed (or “dashed”) onto drivers and their passengers. Now we splash a bunch of numbers on a screen and call it a dashboard. Irony.

The real goal of a dashboard

A common misconception is that the goal of a dashboard is to answer all questions. That is flawed because you assume there’s a world in which a dashboard has the nuance, context, and cross function connective tissue required to truly answer complex questions. It doesn’t.

Example

Revenue is down -9.3% to plan. Naturally, you ask why.

1. So you build a subsequent chart that breaks revenue down into “# of active customers” vs “Average order value”. You see that Average Order Value (AOV) is down -9.3%.

2. So you build a subsequent chart that breaks down AOV into “# of items” and “average value per item”. You see that # of items is down -9.3%

3. And on. And on. And on. This is why every company falls into the dashboard trap.

Instead, the goal of dashboards are to prompt questions that lead to analyses.

Here’s a better approach:

Revenue is down -9.3% to plan. Naturally, you ask why.

1. You spend a few days digging into why it happened this past week

2. Then you spend a few days digging into how long it’s been happening and where the issue is

3. Then you put together an analysis that says the # of items per order has steadily been decreasing for a few months

4. You bring together a few people and you take action

You might look at this example and say “a really good dashboard could have done this”. Fine, maybe, for my one example, but there is no chance that a dashboard can answer all the permutations of questions one has about a business. And if you do want them to answer all the permutations then you will create that many dashboards and hence the trap of 10,000 dashboards.

There’s a better way.

Why you need a dashboard strategy

Anytime anyone hears the word strategy, they immediately jump to this image of a McKinsey consultant adding 10 thousand slides but adding zero value. That’s not strategy. That’s a poor excuse for strategy.

Strategy, when done right, can move mountains, but I’m not here to convince you that strategy is important. I’m here to convince you that having a dashboard strategy is essential.

After 15 years in Marketing Data Science, I’ve concluded that every team only needs 3 dashboards. Anything more and you will see diminishing returns. You might think to yourself, wow that’s a lot of dashboards depending on the number of teams. I didn’t say 3 UNIQUE dashboards. I said 3 dashboards. The scale and efficiency of dashboards come from having many people look at a dashboard repeatedly.

To build impactful dashboards, you must make a tradeoff between context and virality.

Context → The amount of information provided in the dashboard to make an educated decision or estimate

Virality → The amount of eyeballs on a dashboard

You can’t have high context and high virality. It doesn’t exist.

But you don’t want only high visibility / low context dashboards or vice versa. You need a balance.

The Executive. The Tracker. The Optimizer.

Every dashboard must answer a fundamental question. For any employee, there are 3 fundamental questions:

1. How is the business doing?

2. How is my team doing?

3. What should we work on next?

Each of these questions has a score on the context and virality matrix.

We then attach a dashboard name to each of these types of dashboard.

1. How is the business doing? → The Executive

2. How is my team doing? → The Tracker

3. What should we work on next? → The Optimizer

You can think of them as titles for Jason Statham movies.

Now, I know you’re all eager to learn more about these 3 dashboards, but I’ll need just 30 more seconds of your patience as I explain the concept of a dashboard brief.

A dashboard brief

A dashboard brief is like a marketing campaign brief, but for building a dashboard. It should include:

1. objective of dashboard

2. why alternatives don’t work

3. metrics to include

4. how the dashboard will be used

5. expected SLA

6. priority over other dashboards and other data projects

The sole act of forcing people to write out a brief will eliminate 75%+ of dashboards, because people are inherently lazy. For each dashboard below, in addition to explaining it, I’m going to write the brief so you understand why the dashboard exists.

The Executive

The Executive dashboard has one purpose. It’s a global source of truth.

The dashboard allows for anyone at your company to know what executives believe is happening and therefore it forces alignment. This dashboard must act as a source of truth for:

1. Business performance

2. Definitions of metrics

You remove ambiguity which is often a killer for dashboards. The first time someone questions a dashboard to say “Yeah but my numbers says this” the conversation is completely useless from that point forward.

Here’s how to brief “The Executive”:

Do not put goals here. Goals simply add another layer of complexity where people will ask “why”. If you need goals then make a separate dashboard with the same data and add goals. Each dashboard must have it’s own objective.

Example

I wrote a viral post on LinkedIn about Summary which was Uber’s Executive dashboard. Pretty rare for a post about dashboards…

The dashboard was simple (image below):

• 8 high level metrics

• Preset look back windows

• Drill downs by 1 geo at a time

• Growth in absolute and % terms

That’s it, but talk to anyone from Uber and they still love that dashboard. As someone commented on the post the biggest benefit was “Time to value for me - and the ability to gather insights at varying levels of detail. The questions didn't necessarily stop there but it equipped us to have exec and detailed conversations quickly. “

Notice they said “detailed conversations” and not that they made “detailed decisions”.

Low context. High visibility.

The Tracker

The tracker moves one level deeper to a functional level. Here, you should take a subset of the metrics from “The Executive” that apply to your function (think different product teams or different marketing teams) and add targets or forecasts. Adding targets or forecasts is a required step because it adds more context to the dashboard.

Note: If you want to learn how to build better targets, click here.

If all you wanted were the same metrics as “The Executive” but less of them, you’re being a bit lazy. If everyone took a different view of the “Executive” dashboard then you’d replicate the dashboard into 1000x different dashboards.

Instead, “The Tracker” adds more context but reduces virality.

Here’s how to brief “The Tracker”:

This dashboard should be used to align priorities at the beginning of a work cycle (week, sprint, etc.) . Again it doesn’t tell you how to fix the problem but what problems are most important to fix.

Example

Sales vs Targets is a classic example. You simply plot the key metrics for your sales team against a target / forecast and the gap (or lack thereof).

A sales lead doesn’t need to know exactly why they are behind targets from this dashboard, but they absolutely need to know that they are indeed behind targets.

The Optimizer

The Optimizer is the minimum viable amount of data required to identify where to dive deeper. It removes the distractions of targets and too many metrics. It’s designed to narrow your focus but at the same time do some of the heavy lifting for you.

Here’s how to brief The Optimizer:

Example

An example of this is a funnel dashboard . It’s highly contextual.

This is where you can customize to the specificity of teams, but try not to go too crazy.

If you’re an email marketing team, this is a good area for open, unsub, and click rates by campaign. You could also add # of users targeted and many other metrics here. It’s okay if this dashboard gets a bit messy, but don’t add data for the sake of adding data.

Ask yourself: Have I used this metric more than 2x a week for the past 6 weeks? If yes add it. If it’s “kinda” then don’t add it. Trust me.

You should be looking at the optimizer often 1x a week and better understanding your customers behavioral trends.

Wrapping it up

You need a dashboard strategy so you don’t succumb to a death by a thousand dashboards.

The dashboard org chart helps you build an effective framework that balances impact + repeatability of dashboards. Pair that with a well written dashboard brief and soon you’ll enter dashboard nirvana.

Enjoyed reading this?

Share this with your colleagues or on your LinkedIn . It helps the newsletter tremendously and is much appreciated!

Missed my last article?

Don’t be sad. Here it is: How to build better Marketing targets