Conversion Crazed

Social Conscience Cont. Banishing Trans Fats

See what did I tell ya, social conscience: Starbucks Drops Trans-fats

Before you go skeptical on me think about it... with nearly 40 MILLION!! customers weekly this could have an impact.

Metrics and Dashboards: Everything is Illuminated

Last week I was interviewed for B2B Magazine about one of my favorite subjects, Metrics and Dashboards. As with all interviews one never knows how much they will use and hardly ever get a chance to clarify what one has said but it was a good article overall.

I would like to spend a few minutes talking about this subject of Metrics and Dashboards.A properly constructed dashboard with well thought out metrics is a beautiful thing.

There is a really interesting book and later movie called Everything is Illuminated and the most poignant line in both is when the main character has an epiphany that "Everything is Illuminated in the light of the past" (hence the name of the book), not exactly a mind shattering statement but very easy to relate to. In general, this is sort of what a good dashboard does it allows a business to get a clear understanding of the past, understand where they are at the moment compared to where they think they should be, make informed decisions on how to affect those metrics, and one of the most powerful function of dashboards, to help determine causality. It's great to know where you stand and what the trends are but the real value becomes knowing why trends change, what is causing the change. In the article I tried to explain an example of this but a sentence was left out so it probably didn't make sense.

Here is a really rudimentary example of this; Let's say you are looking at your overall paid search channel metrics and impressions are flat, sales are down, and your efficiency has gone down. Several things could be happening here: your average order value could have tanked or your conversion rate could have tanked or your click though rate is down or your cost per click has gone up or it could be a combination or all of those things. The point is that if your dashboard has each of those metrics trended on the same scales you should be able to see the cause very quickly and know where you need to dig deeper and focus your efforts.

After years of creating dashboards I have become pretty familiar with what it takes so I thought I would share some tips/things to think about when creating you dashboards:

1) Determine the audience for your dashboard: One of the first mistakes I see a lot of businesses make is they try to make their dashboards be everything to all people. The result is that it's too detailed for the management of the company but not detailed enough for those in operational roles . This creates a lose/lose scenario where the dashboard is more frustrating than illuminating, and people will not use it.

2) Identify the metrics for your business and the dashboard. There are general metrics that everyone should have (revenue, orders, daily unique visitors, conversion, etc., etc.) but there may be metrics that are specific for the business or the area of the business. For example if you are a subscription business you had better have "retention rate" and it's relative the "churn rate" in your dashboard.

3) Find comparables in metrics. Remember one of the powerful things of a dashboard is it's ability to show you the possible causes of trends. Think about the questions that might be raised by increases or decreases in the trend of a certain metric. Is there another metric that could be added that would allow you to identify or dismiss a possible answer to the initial metrics plan.

4) Keep it simple. This is where I contradict myself. While it is good to add comparables it is really easy to over-complicate and have your dashboard be cluttered. Focus on what is really matters qualitatively and quantitatively. Information overload will cause your dashboard to be the last things anyone finds value in.

5) Have Good Data Hygiene. garbage in.. garbage out. It's so critical that you trust that the data you are making decisions, or gaging your performance from, is reliable.

7) Make it easy to update. Bottom line if it's not easy it will never be updated. Try to automate as much as possible for two reasons: First, you want to make sure you are spending time getting value out of the numbers not compiling them. Second, if it's manual the opportunity for user error is large and nothing is as scary as making a big decision based on bad metrics.

8) Think about how it's distributed. Are you distributing it in a way that causes it to quickly fall to the bottom of a large stack of papers on someone's desk? Is it lost in a jumble of emails that everyone gets first thing in the morning? Are you reviewing it as a team? Reviewing metrics is definitely one of those times where two heads are better than one.

9) Consider at what frequency it should be updated. Frequency is important but update it too frequently and the changes are not meaningful, wait too long and you miss the opportunity to head off a problem or take advantage of an opportunity. Also the longer the wait the more information you will be forces to evaluate.

10) Decide who owns what. Sounds unimportant but at the end of the day someone has to be responsible for each or all metrics both from a data reliability and from a decision making standpoint. 

11) Layout and looks are very important. I cannot emphasize this enough. I have seen great data completely destroyed by bad presentation. If it looks good and is easy to interpret it will be looked at and valued. Don't require a magnifying glass. Don't feel the need to report the number for every instance in the trend when the trend itself will paint the picture. Use graphics, use color. Use pie charts sparingly as they are a moment in time and not a reflection of change. Overall be esthetically pleasing.

There are a lot of people thinking about this right now but one of the best thought leaders on how to communicate data effectively is Edward Tufte, I'm a big fan!. Specifically he has coined a style of reporting called sparklines. While this is relatively new I doubt anyone hasn't been exposed to sparklines in sport scores, stock reports, news stats, etc. He has a great running post located here (caution the conversation gets really convoluted and philosophical at times, concentrate on main subject) Part of the discussion reviews a piece of software that I use in my dashboards called MicroCharts by BonaVista software. Here is a screenshot and PDF of the type of dash you can make using it.

PDF Example

There are lots of others ideas and I could spend all day talking about metrics and dashboards, but these are some of basic to get you started. I know that monitoring metrics and using dashboards can make a huge difference for any business willing to take the time to make sure they are looking at the right things. If you have good tips or tricks, ideas, etc. please share.

Business Innovation: Prevent the Ordinary

I've been thinking a lot about business innovation lately, and all around the idea of how long can a company ride on the coat tails of a truly brilliant innovation. I have been thinking that lately, in our more innovative and fast paced market, the time one can ride on previous innovation is less and less.

It was awhile ago but I have to belatedly say hats off (once again) to John Moore at Brand Autopsy for it string about Harvesting Collective Genius. Part I was great and part 2 a few weeks later was no slouch. The videos posted from Business Innovation Factory's Collaborative Innovation Summit #1 were great. I particularly enjoyed Steven Denning's comments on creating demand for innovative ideas through a compelling narrative story. He uses his example of changing the World Bank to be a knowledge sharing organization through interesting storytelling patterns. Steve provides a great method for you to cause innovation to happen in your organization in order stay ahead of the curve.

See the video of the presentation

More info on Steve Denning