Smart Dashboard Modules: Insightful Dimensions And Best Metrics

red sparksMy last post, perhaps provocatively, called for a reduction of data in executive dashboards (digital, online, offline). More English (IABI, specifically) would lead to a smarter understanding of performance, and of course glory for data practitioners.

Here's the post: Strategic & Tactical Dashboards: Best Practices, Examples.

In the post Adil commented that he's observed that attribution modeling is missing from most web analytics dashboards. My reply was that attribution modeling might not be present in a physical manner, but that it should be there below the surface to ensure our executives have a smarter understanding of a channel's true contribution.

While that is a perfectly good reply, Adil did make me wonder what a perfect dashboard module (one element of a dashboard) might look like should we want to parade the dead body of last-click attribution in front of our executives.

I'll share the outcome of that thinking in this post, but you also get a bonus. The post also shares a total of seven complete modules to your digital analytics dashboards. Each shows a unique facet of digital performance, and I've attempted to choose the best dimensions and, for those dimensions, the best possible metrics.

If you are in the process of creating a dashboard, you'll discover that this post saves you a million hours of wandering in the woods (ok, so not a conservative estimate!). If you already have a dashboard in place, the post should spark introspection about your choices and encourage a few valuable changes. The post shares seven distinct, awesome, dashboard modules.

Before we go on….

Prologue:

You'll recall the killer insight from the post on strategic and tactical dashboards, getting the balance between data and English right, and the shift involved as you move from customized data pukes (CDPs) to tactical dashboards to truly strategic dashboards…

best dashboards for each management layer data english words 11

This post is focused on helping you create smarter tactical dashboards. The target audience would be Senior Directors and VPs in the company/division. They will have some of data, but your insights, recommended actions, and computation of business impact (IABI) will be critical in driving change.

Creating these tactical dashboards (I'll end with a representation of what an entire dashboard could look like), will give you much needed experience to step up to creating magnificent strategic dashboards.

Before we go on, what do I mean by database module? It is one complete representation of performance, usually illustrated in a table or a graph. Here's one simple example that is likely in every single digital analytics dashboard…

sessions conversions google analytics 1

The dark blue line is Sessions (Visits, if you are not using Google Analytics) and in the light blue line is Conversion Rate. Ever single dashboard likely has this module because every executive wants to know how many visits were there on our site, and what's our conversion rate!

This is an easy module to create in Google Analytics (or any other tool). Most of the time it is value-deficient, but sometimes, like above, it does raise interesting questions. If we are getting so much better at getting traffic, how come our conversion rate is essentially flat? If we've spent money on qualified traffic, why is the light blue line not sloping up over time? Good question!

Oh, and this is why you need the English part (IABI). Without it, you are assuming the dashboard recipient can answer that question and that is just silly.

Ready to go? We are going to focus on tactical dashboards, and here are seven delicious dashboard modules I've created that should help you create insightful dashboards. I've done the hard work on your behalf to pick the dimensions and critical few metrics that will illuminate your complete business performance. [And Adil, if you are reading this, #2 is just for you!]

Ok, don't forget to breathe, this gets exciting!!

1. Search Contribution: Acquisition, Behavior, Outcomes!

Before we get to attribution modeling here's one of the more basic dashboard module, and a best practice.

Search engines are immensely valuable sources of traffic for any type of website. For the site below, they represent 56% of the site traffic (in the best practice range). So obviously the executives will want a dashboard module that represents search performance.

Given our challenges with not provided, (See: Search: Not Provided: What Remains, Keyword Data Options, the Future), it might be optimal to just have the search engine level view in the dashboard.

Here's the best practice I encourage you to consider… show the search source and Users, Pages/Session and Per Session Goal value.

search engine effectiveness 1

How many people visited our site, during their visit how much content did they consume, and what was the value of the macro and micro outcomes to our company?

There are a million other things you can add to effectively confuse your executives. You might not need to. This tells them everything they need to know.

A big, large, huge, gigantic, mistake Analysts make in dashboard modules is they only show a partial view, rather then end-to-end. So for Search they might show Users, Sessions, CTR, CPC (for paid), Impressions, and Conversion. The problem is they over-indexed in acquisition (the first five metrics), only had one outcome metric, and no behavior metric. Always focus on the complete journey! (See: Best Metrics For Digital Marketing: Rock Your Own And Rent Strategies).

In our module above we have one of each to represent the end-to-end journey so that you can judge which engine is bringing lots of traffic, but then is it traffic that engages and in context of that do they deliver value to us? Data above is for a B2B website with no ecommerce. Always compute economic value!

As you can see my dashboard module is directly from Google Analytics. You can use the standard Channels report, or you can create a custom report. I'll show just the standard reports directly from GA for rest of this post. But you can also use the free Google Analytics API to automate extraction of the dashboard modules and create pretty dashboards. There are many apps in the Google Analytics App Gallery that also do this very nicely.

With a tiny bit of love I can take my slightly ugly, but still insightful, dashboard module above and make this pretty thing…

search dashboard module 1

Nicer fonts, some simple application of conditional formatting (try it in Excel), and boom! Your boss loves you. Oh, and yes, it is much easier to see patterns in the data. I suppose that is a bonus. :)

For the rest of the post, I'll just share the standard GA view because I just want to create the modules and share the best dimensions and metrics to focus on.

2. Channel Contribution: Assists AND Last-Click Impact

One thousand, one hundred and three words later, we are finally on to our friend Adil's comment that inspired this post.

There are three types of multi-channel attribution scenarios. Almost always your efforts to apply custom or standard attribution models will show up in tables and graphs that measure conversions/outcomes. They will be silent and behind the scenes.

But this module could sit all by itself on your dashboard and both educate your senior-most executives and get them to ask all the right questions about the thing you want from attribution modeling efforts: Are you investing our marketing budget optimally in our media-mix?

On the left are all your owned, earned and paid media channels. Then we show the number of assisted conversions driven by the channel. This is super-duper important. People came to your website (mobile or desktop) and they did not convert during the first visit (that is absolutely ok!). But they did come back in the future (in a 90-day period) and convert via a different channel (also absolutely ok!). For the first time on your dashboard you'll give these beautiful channels credit they deserve!

channel contribution multi channel funnels view 1

The third column is what your digital analytics tool will report, how many last-click conversions did the channel deliver. This, of course, is important. It is just not the full picture.

The very last column is to deliver a quick sense for your executives for which channels tend to have more propensity to deliver last-click conversions (see Paid Search, Display), and which might introduce just as people people to your business (assist) and convert about the same (Social, Email).

This dashboard module will cause your executives rethink ad-targeting, ad-content and landing page strategies (much more selly, sell, sell on Paid search, a lot less on Email and Referrals). It will also encourage them, without having to understand the immense complexity of attribution models or media-mix modeling, to give you permission to invest in aforementioned complexity.

Win-win. Achieved simply.

3. Content Contribution: Consumption, First Impressions, Value Delivered

Changing gears, you'll surely have one database module to communicate the value of the content your company is investing in. Plain text. Product recommender tools. Configurators. All the lovely See and Think content. Awesome videos (boring videos). Pimpy Flash thingys (remember… every time you use flash on your website, a puppy dies).

Here's my favorite content contribution dashboard module…

content contribution multi channel funnels view

Several things to note here.

Do you see that I cover acquisition, behavior and outcomes again? Always do this!

Did you notice I'm biased towards measuring people wherever I can, I don't like sessions/visits as much (unless I need to align users, sessions, and hits optimally).

Unique Pageviews only counts the page's view just once, even if a person viewed it ten times. Entrances is there to give us a sense for how often the page is a landing page (remember… every page on your website is a home page!). Finally, my bff Page Value shows how much money a give page made for us (it works both for ecommerce and non-ecommerce websites). You want our executives to move away from obsessing about page views, to valuing outcomes. Look at the wide spread above, from pages making us $0.27 to $2.75!

How many people saw the page, how many entered on the page, how much money did it make? What else do you need to know when it comes to content consumption if you are an executive?

This module will always provoke questions (be ready with your IABI) and drive short and long term change.

4. Country Contribution: Intensity of Outcomes, Something Big, Something Small

Executives don't worry about geographic differences enough. Cultures. Spending power. Shopping behavior. To name just a few.

In my dashboards for companies that have a wide geographic presence, I like to include this dashboard module…

country contribution micro macro conversions 1

In this case we are focusing on countries, makes the most sense for this business, but you can just as well do cities or DMAs or a different more optimal slice.

Because this is just a couple times a year type business (at best, mostly you might engage just once a year), we show Sessions (I want to show you that you should read best practices, like mine in the above module, but then always, always, apply local knowledge of the business).

Then the difference in Goal Value between each country. The Comparison button on top of every GA table helps us eliminate most numbers and replace them with this red and green indexed value against the overall average.

And finally, one micro-outcome. Geographic differences often drive different outcomes. The Per Session Goal Value covers all the goals and outcomes. But I wanted to call out Goal 2 as a proxy for differing behavior. Again, using the Comparison feature. In your dashboard module, customize this column.

To create this report in GA, go to Audience > Geo > Location > Click Comparison on top of table, choose the right metric (Per Session Goal Value and Goal 2 used above).

Get your company to be smarter and more targeted about their advertising and marketing strategy by using a geo segmentation. Your shareholders will thank you.

5. Social Contribution: Activity + Full Credit For Outcomes.

With all the hype around Social Media (yes, still!), can you imagine how many things we can show in our Social dashboard module? Total Likes! Total Followers! Circle Members! Pageviews! Wait. What was that last one? :)

Ok, there is a lot.

I normally show a module that contains my best social media metrics: Conversation Rate Amplification Rate, Applause Rate, and Economic Value.

But recently I've gone simpler, partly to accommodate for the fact that our executives have become so much more skeptical about all the social hype and its value. Especially, executives that own outcome oriented sites.

Here's the dashboard module I'm using a lot more, and the response from the leadership has been pretty ecstatic.

If you go to Acquisition > Social > Overview, you'll see it.

The small blue circle is what our executives are often told Social channels delivered as outcomes and economic value. (Data below is for a non-ecommerce, B2B website.)

The light blue circle is what they should really be told, outcomes and revenue is roughly 2x of what they believe. We want to give Social credit for all the non-last-click conversions! It is only fair.

social value module

The big circle represents overall outcomes. 315k conversions. The two numbers at the bottom, orange box, are there to provide important context.

Now our executives have the full picture. Social is important, but it is not the thing to most obsess about. If you want only short-term outcomes, inside 90-days, invest less than fifty nine thousand dollars (people, process, advertising) in it and you'll break-even. If you can afford to be patient and believe in long-term outcomes, invest more.

A simple database module that gives Social its full credit, and allows for smarter decisions.

Our executives will always have pet projects, they will always have the next shiny object to chase. Your job is to create the simplest database module for that shiny hyped thing so that you can help the executives get to the best decision.

6. Product Contribution: Bestsellers, Crazy Buyers, Opportunities.

Ok, enough of all this B2B and non-ecommerce stuff.

Let's focus on money! Ecommerce! Life-sustaining ecommerce!!

This might surprise you, but my favorite ecommerce dashboard module shows product performance…

ecommerce dashboard module

We obsess at high level conversion rate type stuff so much that we don't often look at what is being sold. And even if we do we might list the total quantity of a product ordered and it will take the shape of a top ten products sold dashboard module. All fine. It is not insightful enough.

My favorite metrics are Unique Purchases (total number of times a specified product – or set of products – was a part of a transaction), Average Quantity (average number of products sold per transaction) and Product Revenue (revenue from individual product sales).

Take a couple minutes. Look at the table. So weird right?

Why are people ordering, on average, 2,822 quantity of the toddler tee? What do we need to learn more about the product, or, better, the purchasers? The Fiji pens look a lot more normal. But look at the motion lamp or the messenger bag. Weird stuff going on!

That is why I like this module, structured this way. It brings the data to our executives to change minds about what our customers are buying and how much and forces us to ask why. And as the best analyst knows, asking why always leads to good things!

Use this module. Surprise your company stake holders.

7. Intelligence Alerts: Greatest Starting Points, Brought Forth.

My last dashboard module falls is from Planet Wish-list. It is less something you can get directly out of Google Analytics, like all of the above modules, and more of something I would love for you to spend time and money on (usually with a consultant).

Most dashboards don't obsess about what's changed, they focus more on what happened. And that's not the same thing.

The Intelligence Alerts feature in Google Analytics does focus on what's changed. In English… here are some things in your data that are showing an unusual pattern,let us present them to you in the order of importance (statistical significance).

It is really cool. If you go to Intelligence Events > Overview, you'll see something like this…

intelligence alerts importance google analytics 1

It shows the impacted metric, the segment where the behavior was observed and the observed change and how much confidence does Google Analytics have that the change was important/significant.

You can click on any of the alerts above and you'll get more detail….

intelligence alerts deep dive

In this case you see the expected range of average session duration (maxing out at one min and four seconds), and the actual performance was almost five minutes!

Does your dashboard have a module that represents this type of change? Give your executives the best starting points for their questions? It is important to illustrate what happened as you see it in all the other dashboard modules above, but I believe the best dashboards in the world, representing the work of the best analysts in the world, will have a module that represents results of automated analysis of the data to highlight at least the known unknowns (and if you are really good, some unknown unknowns!).

In case you are unable to build your own system or module, please leverage the automated intelligence alerts in Google Analytics. And for the known unknowns, do remember to set up relevant custom alerts.

Closing Thoughts.

We should strive for the best strategic dashboards in the world. Our leaders deserve nothing less!

But, it is of value to create great tactical dashboards for people in our company with enough connection to data and reality. The seven modules above cover your entire business, and leverage important frameworks in their presentation (critical few metrics, obsessing about people, focusing on acquisition, behavior and outcomes).

When it comes time to pull it all together, here's my recommendation for what your dashboard could look like…

digital analtyics dashboard example sm

A dashboard module on the left, a text box with the insights, actions and business impact (IABI) on the right. Please click on the above image for a higher resolution version.

Your executives will be informed, advised and will take every action you recommend, and then some more.

Life will be… glorious.

I wish you all the very best!

As always, it is your turn now.

Which of the seven modules is your favorite? Is there one that you tend to use across multiple companies/divisions? What does your favorite dashboard module, from any too, contain? Which facet of a digital business do the above modules ignore? There are lots of compromises in picking the critical few metrics, which compromise goes too far? What else would you add to fix it?

Please take a moment to share to your delightful comments, insightful suggestions, valuable examples via the comment form below.

Thank you.

Comments

  1. 1
    Petros Z says:

    That is a very insightful post.

    Getting to the meaning of endless data, should be a responsibility shared among the producer of the dashboard and the reader. A person that is close to the metrics should give insight but also not choose what kind of information to 'hide' because he considers it not important.

    Many times, I push myself to decide the metrics I want to check before I actually get them. This is because it is easy to get sidetracked by an anomaly in data and actually miss your initial point.

    Thanks for the wealth of knowledge and expertise you are sharing here.

    • 2

      Petros: I concur with you on the value of having a partnership between data consumers and creators. I believe the Digital Marketing and Measurement Model can be a good aid in fostering that discussion and getting each party to do their part.

      To your point on "hiding," :), there is always a healthy tension between what executives want and what they really need. At the start of the relationship I will have 90 want and 10% need (even on day one!), and then slowly over time we can get them to just 20% of what they want and 80% of what they need because we would have earned their trust and respect for our skills.

      Thank you for sharing your comment, and wisdom with the readers of the blog.

      Avinash.

  2. 3
    Shiv Gupta says:

    Thanks for this post.

    Currently I am working for an ecommerce site and I have increased it's traffic and sales but I am not able to analyse the last click and first click.

    Please provide any valuable source..

    • 4

      Shiv: Any modern enterprise web analytics tool will allow you to do attribution modeling. Google Analytics is a free option you can try.

      You mention first-click, I actually highly recommend against that attribution model. For more on that, and other valuable models, please see this post: Multi-Channel Attribution Modeling: The Good, Bad and Ugly Models

      All the best!

      -Avinash.

      • 5
        Shiv Gupta says:

        Thanks Avinash Jee..

      • 6

        I favour first-click attribution when work for clients in the travel industry. That is, the first referral source gets a significant weight over the others.

        The theory is, in an industry with so many competitors, the initial lead has the highest value; subsequent follow-up visits cement the relationship; by the time of the last click, the decision has already been made.

        The thing is with attribution models – all are right and all are wrong! Its just that some are more right and less wrong than others…

        • 7

          Brian: I completely agree with you that unless one goes down the path of true controlled experiments and leverage true media-mix modeling, attribution modeling is a path down just being less wrong. How less wrong you want to be is based on our research/assumptions.

          In my best practice custom attribution model, first click does get some credit. But like all other subsequent touch-points, it has to earn more than the some credit. The second half of the model is the "multiplier" so, if as you say, it is actually valuable, it can earn a lot more credit.

          But I've been personally unwilling to, even in travel type scenarios, to simply give it a lot of credit without demonstrating value. Based on your experience I'll be on a lookout for how less wrong this strategy is. :)

          Avinash.

  3. 8
    jose leite says:

    Mega post Avinash, thanks :)

  4. 9
    Indu says:

    Hi Avinash – my favourite would be Channel Contribution: Assists AND Last-Click Impact.

    We are working on retargeting as part of a campaign for the client. Since there are multiple business owners with accountabilities across different regions approvals can take forever with conversations going in circles.

    Something like this would be the step in the right direction in making a case to the client.

    Thanks Indu

  5. 10

    Always a pleasure to gain knowledge from the best, in an ever changing world. or as Socrates once said,

    "To know, is to know that you know nothing. That is the meaning of true knowledge."

    • 11

      Eric: That is meta! Thank you for sharing my new favorite quote. :)

      There is so much changing in this field, and the boundaries are ever expanding. I'm reminded of my other favorite quote: Stay hungry, stay foolish.

      Avinash.

  6. 12
    Chip says:

    Avinash,

    I think these are all really great examples of what you could/should be reporting to C-level executives. Like you always say "Segment or die and ask the same three questions: So what?"

    Do you think that segmenting your Acquisition, Behaviors and Outcomes based on User Type would give your dashboard more or less insight? In this case of high-level metrics, would it matter?

    • 13

      Chip: Hmm… that is a good question.

      What do you mean by User Type? Google Analytics defines it simply as New and Returning Visitors, I think that is totally useless. But perhaps you have a different definition. If you share it, I'll be delighted to think some more and share my humble perspective.

      Thank you,

      Avinash.

      • 14
        Chip says:

        Yes, New vs Returning. With most of our e-commerce sites and non e-commerce with goals, I've found a considerable difference between goal completion rate and e-commerce conversion rate between new vs returning users.

        I think that data coupled with frequency & recency reports could be helpful.

  7. 15
    Adil says:

    Hi Avinash,

    Thanks for taking the time out for this post. Totally appreciate it. I think that by bringing the MCF to the forefront via a dashboard, analysts can get high level questions going from Executives – understanding the assists/last click ratios >> leading to attribution modeling to dig out more and helping with media mix. Everybody wins.

    For me, this particular recomendation ties in well with your recommendation #5: Social Contribution, as all executives have an eye on social right now. Not only does the MCF show that social works by assisting (hopefully), what would be the $ value on the assists >> Meaning that comapny X is doing something right on social with regard to See +Think part (if not Do as well).

  8. 16
    David Wells says:

    Doesn't "Why are people ordering, on average, 2,822 quantity of the toddler tee? " on point 6 seem like an error in the stat reporting?

    Also, 8 unique purchases of a 20oz Motion lamp with the average order quantity of 3751?

    Seems odd =P

    I'm wondering how often, these numbers are investigated when you see an outlier like that.

    • 17

      David: I neglected (ok, on purpose did not) to mention the website the data is from. It is not entirely usual for them. Maybe just a little.

      But my point is that I rarely see this type of stuff on dashboards (it could be sampling bias). Hence, I'm making the case that don't be satisfied with overall revenue and just top products sold by quantity. Have something that digs a bit deeper and highlights this type of behavior. And if you find it, to follow your guidance, investigate!

      Avinash.

  9. 18
    Jason Juan says:

    Hi Avinash,

    I'd like to share 3 thoughts with you.

    1. One of my favorite dashboard is your no6 Product Contribution. However, I usually use Category contribution instead. I like to compare the traffic (visit) and outcome (usually sales units) and I can very quickly figure out what categories are my "traffic driver" and what categories are my "sales generator". It could be consolidated with category management (we can choose the category captain based on that data)

    2. Another my favorite dashboard is new vs. returning visitors. When speaking to marketing spending, we will focus more or less on acquisition or retention. Hence, from the proportion of returning visitors across time spans, we can understand if our acquisition/retention takes effect or not.

    3. For all the dashboard, I use comparison metrics (MOM or YOY) to help me judge what happened by now. Especially I like YTD(Year to date) YOY, it eliminates the seasonal and promotional effect and give me a clear view about how far we're ahead or behind.

    Hopefully you will find the above interesting.

    Jason

    • 19

      Jason: Love your thoughts!

      Could not agree more with you on #1 and #3. Excellent advice.

      I have to admit I actively recommend against new and returning visitors to our Master Certification students at Market Motive. Especially returning. What is that? Someone who came 10 times is returning, as is someone who came twice. With a lack of clarity, you can't actually find decent insights.

      I recommend two things.

      1. If you are an ecommerce type site, go to the Path Length report, understand the distribution there and then make a more informed choice. "Visitors who came more than four times because on average people convert after five visits." Just as an example.

      2. If you are a non-ecommerce type site, look at the Recency/Frequency reports and identify what numbers make the most business profitable sense and would also mean delivering value to the user. Then pick that number for your segment, apply it to your report, rejoice!

      In both cases you measure a specific behavior, it helps us make smarter decisions. Does this make sense?

      Avinash.

      • 20
        Jason Juan says:

        Hi Avinash,

        Thanks for your response. I'm glad you like my thoughts.

        For the path length report you mentioned, what could be the solid recommendation or action to business people ? For example, if I know people who came more than 4 times will tend to convert, how can I encourage them to visit more ?

        Well, I know it's a big question but just want to have some of your inputs !!

        Thanks,
        Jason

        • 21

          Jason: You can change your behavior and the incentives you give them. That is a very broad umbrella, so let me give you two examples.

          You can change your ad content and landing pages for long path length. You know it is going to take them a while to convert, make sure your ad content and landing pages are not pushy. On landing pages rather than just focusing on product features, give them more information they might need to make a more considered decision. Things like that.

          On the incentive side… if we know that you are not going to covert on the first visit on, say, our travel site, then at the end of you searching for flights you see a Buy button but also a Save Your Results button because most people won't buy. But by getting them to save results we are going to get their email id (to send reminders, with permission) and next time they come back we will have saved them more work. More things like that.

          One advanced use of Path Length and Time Lag data is to use it for behavior targeting with an advanced personalization engine. It is hard and sometimes expensive so I won't go too deep into it, but I want to share that that is an option.

          Avinash.

          • 22
            Jason Juan says:

            Hi Avinash,

            Thanks for the advice. It's clear and helpful.

            Usually we get some insights but just don't have ideas on how to transform the insights into actions.

            Your advice are very good examples. I really learn a lot from you.

            Thank you,
            Jason

  10. 23

    This is absolutely brilliant Avinash.

    We have started to use the GA Embed API to produce dashboards and encourage insights to be placed alongside the charts / data tables, so I love adding Actions and Business Insight to this.

    Also, I love the CDP – made me laugh!!

    Do you come to the UK very often? Just wondering if there are any events you are speaking at, or indeed whether if you were whether you would be interested at speaking at some of our internal events.

    M

  11. 24
    Kevin says:

    Hello Avinash,

    Great stuff as always!

    Do we have any insight as to where Google Analytics gets the expected ranges for intelligence reports?

    I imagine they use historical data, and current rates (i.e. conversion rate for eCommerce statistics) to create a standard deviation, but I'm just guessing.

    Also, in that same realm, do we know what determines the level of importance for the noticed change? Have they revealed any more details on this, or are we kind of in the dark for this unfortunately?

    I did a quick search for more information on it, and couldn't find anything of note. Any help is appreciated. Thanks!

  12. 26
    Varun Sarin says:

    Hi Avinash,

    Thought provoking post as always- what really jumps out is the amount of work many organisations have to do to even begin capturing the data behind these dashboards. Atleast mine does.

    I have a question on your content contribution dashboard (#3) – Would it be too confusing to add conversation and amplification? This would help visualise how our content marketing efforts are resonating with customers. Or would you capture that as part of the social contribution dashboard instead (#5)?

    Thanks,
    Varun

    • 27

      Varun: For a dashboard module you want to be super-duper focused. So broadly speaking adding any more metrics to any of the modules in this post would be sub-optimal.

      But. We can absolutely customize each module for your business uniqueness. If I am creating a content contribution module for this blog, I might choose Unique Pageviews, Conversation Rate and Page Value.

      I love amplification rate (http://zqi.me/rocksocial), but I would skip it here. Though there are other cases where I might eliminate one of the above three and go with Amplification Rate (for example, in the analysis of my Facebook or YouTube efforts).

      -Avinash.

  13. 28
    Adam K. says:

    Hi Avinash,

    Another great post! I'm relatively new to the world of data analytics (I've only recently climbed out of the pit of 'data entry' *shivers*) and your posts are really helping me gain some perspective right off the bat. I'm hoping that they are helping me to appear wise beyond my years.

    One thing I noticed was absent here was an out-and-out engagement metric. I'm mainly concerned with CRM in my business and therefore don't use loads of Google Analytics in my day to day work, but I'm sure you could create a goal along the lines of "signs up to news letter – at homepage". I use something similar in my weekly report and it's great (the "heads of" love it!) showing numbers of people who have signed up to our mailing list through different channels. That paired with their contact status (mailable, e-mailable, both, neither) makes for some great insights about how well our customers are connecting with us as a brand (i.e. not just stuff they buy). All of those figures are of course compared to "this time last year" to give that all important change view on what's changed.

    Currently your dashboard tells us the following about the customers:

    1. Where did they come from? (which country, channel, or search engine?)
    2. What did they do? (conversions or perversions?)

    …but by adding an engagement metric we get this too:

    3. Why will they come back?

    What do you think?

    Cheers,
    Adam

    • 29

      Adam: A part of what you describe is what I call a "micro-outcome" (signups for newsletter). You see it in pretty much every module in this post (contributing to metrics like Page Value, Goal Value etc.), and more specifically in module 4.

      In module one the pages/session metric is the only one in this post that is around "engagement." But an easy search on this blog will share with you some of my other favorite metrics of that type: Visitor Loyalty, Recency, Conversation Rate, and others in that spirit.

      I have to admit I'm not a fan of using the name Engagement for any metric. More on that here: Engagement" Is Not A Metric, It's An Excuse

      Good luck!

      Avinash.

      • 30
        Adam K. says:

        Hi Avinash,

        Thanks for your reply. I had a no idea "engagement" was such a dirty word! Having read "Engagement" Is Not A Metric, It's An Excuse" I realise now that what I'm referring to as the 'e-word' could be simpler (and require far less generalisation).

        Visitor Loyalty and Recency metrics rock, probably much more indicative of the results from well engaged customers than any single action a customer could take. I'll definitely keep that in mind!

        Cheers,
        Adam

  14. 31
    jerome says:

    Nice one…Thanks for sharing this!!!

  15. 32
    Chuck Ullan says:

    Hi Avinash, I love most of the modules here as a periodic analysis, but not necessarily a dashboard.

    It seems like there's a bit of a paradox here in that weekly data pukes are needed to make sure there are no fires and everything is on track, but that same puke can never tell you why.

    For example, module #1 (activity by search engine) is useful, but only periodically. And even then, it's more insightful to see what's changed and why. Week to week or month to month, there are rarely consistent new insights to share for that exact view.

    In my experience, the vast majority of insights come from deep dives in a particular area and would be worthy of their own report.

    For that reason, I love #6 and #7 as it highlights the unusual and new and will probably not look the same.

    Perhaps there is some difference in our definition or expectations of a dashboard. In any case, I love all the analysis examples and learned a couple new things!

    • 33

      Chuck: I'm glad you found the modules to be of value.

      I'll go as far as to say that data pukes, regardless of how much data they contain, will likely never tell the Sr. Executive the way. Because of the challenge at the top of this post.

      We have to do the deep dive and we have to tell them why (I) something is happening, what (A) to do in our best opinion, and how much impact (BI) to expect.

      Perhaps over time our executives trust us so much that they simply as for the IABI and ask us to keep the CDP. :)

      Avinash.

  16. 34
    Pramod Khanchandani says:

    Hi Avinash,

    Excellent post (as always) and very timely for me as well.

    I am wondering why have you not used Bounce rate anywhere as a metric for behavior or acquisition quality.

    Could not thank you much for all that you write… You have been a guide.

    Regards,

    Pramod

    • 35

      Pramod: I do like Bounce Rate as a metric: Standard Metrics Revisited: #3: Bounce Rate

      I do not believe that Bounce Rate is a key performance indicator (KPI). The difference between a metric and a KPI is simply that a change in the KPI directly moves the bottom-line. You can reduce Bounce Rate, remember, an important metric, from 100% to 10% and yet if the rest of the experience is not up to snuff, it changes nothing.

      So, think of Bounce Rate as an important diagnostic metric and unless you have a compelling reason, don't make it a KPI. (And only KPIs typically end up on the dashboard.)

      Avinash.

      • 36
        Pramod Khanchandani says:

        Thank you Avinash for taking time to respond me.
        Your logic makes perfect sense.

        Thanks again!
        Regards,
        Pramod

      • 37
        Rodney Robinson says:

        Thanks for explaining the difference.

        Up to this point, bounce rate has been a kpi in my book. This is good perspective, and serves as a reminder for understanding the purpose of the data that you analyze.

  17. 38

    Thanks for breaking this down!

    This is great insight into analytics and how to apply and understand the metrics and behavior.

  18. 39
    Malcolm Galt says:

    Very well written article, thanks!

  19. 40

    Finally, catching up with your latest posts!

    Adding to this one:

    Enjoying as I do your Glorious Executive (tactical) Dashboard, I am missing some additional structure. I would expect a connection between actions and pre-set goals/goal values (even if tied to strategic KPIs not present here but feeding on this data). As a result, the Business Impact estimation requires loads of manual work with every instance of the dashboard.

    As for the “contribution” module, I think it is a very good way to escape the multichannel attribution black hole (bending space and time is the least of it). Of course, the problem lies in the absence of separate strategic goals for acquisition, behavior and outcomes… I would rather tie entire channels to long term, medium term, and short term objectives when it comes to simplifying.

    Thanks!

    Sergio

    • 41

      Sergio: Totally, absolutely right… For the best strategic dashboards, there will be manual work involved (given the importance of IABI). For tactical dashboards certain degree of automation is possible, for CDPs total automation is ok!

      Thanks for highlighting the importance of having pre-set goals or targets in the individual dashboard modules. Having goals to shoot for, or targets, is definitely the mark of a savvy analytical practice.

      Avinash.

  20. 42
    Harsh Mehra says:

    Excellent post, you have pointed out some great details.

    Our re-imagined dashboard will allow us to transform our digital business via IABI.

  21. 43
    Caiuspupus says:

    Hello Avinash,

    I don't understand where the following data "Assisted conversion" come from. How can you measure offline conversions?

    Thanks.

    • 44

      Caiuspupus: Assisted Conversions metric will come from your web analytics tool. It is computed automatically for you if your analytics tool focuses on people centric analysis.

      In Google Analytics you can see the metric in your Multi-Channel Funnels report. The way it is calculated is that for each visit to the site Google Analytics will keep track of each referrer (Owned, Earned or Paid) and give you a list of campaigns that assist in a future conversion delivered by the same person.

      On your second question, here's a blog post on how to measure offline conversions: Multichannel Analytics: Tracking Offline Conversions. 7 Best Practices

      Avinash.

  22. 45
    Martin says:

    Brilliant article, I'm just starting out and will need to re-read this a few times to understand properly.

    Also, thanks for the links to further support!!

  23. 46
    Kevin says:

    I've been also in the the process of creating a dashboard, and I agreed that I' had discovered that this post saves me a million hours of wandering in the woods.

    Now I should follow to have a dashboard in place, the post should spark introspection about my choices and encourage a few valuable changes. A big help.

Trackbacks

  1. […] Smart Dashboard Modules: Insightful Dimensions, Best Metrics, http://www.kaushik.net […]

  2. […]
    Avinash Kaushik beschreibt in seinem Artikel 7 Dashboard Module, die er entwickelt hat um die jeweils einzigartigen Facetten der digitalen Performance abzubilden. Wie immer ein überagender Beitrag und dazu noch seine 7 Dashboard Module als Zugabe.
    […]

  3. […]
    Smart Dashboard Modules: Insightful Dimensions And Best Metrics – kaushik.net
    […]

  4. […]
    Good data analysts are storytellers. Effective data visualization is often helped with text. Simple headlines or text boxes help explain what the data is saying and the actions that should be taken. This example below from Avinash Kaushik, Digital Marketing Evangelist at Google, shows how data visualization might be presented to the C-Suite.
    […]

  5. […]
    Smart Dashboard Modules: Insightful Dimensions and Best Metrics
    […]

  6. […]
    Smart Dashboard Modules: Insightful Dimensions And Best Metrics
    […]

Add your Perspective

*