Blog

BLOG

Please, Not Another Dashboard

PUBLISHED ON 3 JANUARY, 2018

(Note: this post was originally published on LinkedIn).

I have spent a few years connecting the dots between performance management, data analytics, and “digital” environments (primarily marketing and ecommerce).

This entails referring to “dashboards” in multiple instances, not as the ultimate deliverable, but rather as an embodiment of the compromise between:

1. Highly actionable decision-support tools (the scorecard, or a collection of “Whats”); and
2. A certain level of agility/independence in the exploration of KPI updates (or a collection of quick “Whys”).

Now, all compromises are naturally subject to alternative interests pulling in different directions (I have discussed “delivery vs. analysis” at length), and that is OK. But this one was completely out of control from day one, given the level of popularity that the very term “dashboard” was already enjoying – in connection with data-enriched, visual summaries of all sorts.

Just take a look at the marketing space alone.

While we (at Sweetspot) have been busy focusing on data verbalization (headlines!) and integrating a workflow between analysts and decision-makers (“Digital Insight Management”), others have put the accent on visual discovery (data visualization with exploratory purposes) and self-service analytics. While some focused on executive KPI updates, others aimed for the highly-tactical (“real-time”!).

And that is not even taking into account the “digital” data challenge. While IT and business intelligence (BI) departments insisted on pushing new data sources through their existing, highly-structured, granular data-based pipelines (spending months and years in the process of getting nowhere), marketing executives built and bought their own tools, specifically designed for the challenges of digital data (mostly semi-structured and event-based) – just look at Gartner’s Market Guide for Marketing Dashboards: you will not find a single BI or data visualization tool.

Of course the problem did not stop there: most of the latter were great at handling ad server or PPC feeds but showed a poor understanding of other sources. Digital Analytics tools had to release their own dashboards to do justice to their own data.

In other words, many of us have quickly found ourselves both enjoying the easy label (unavoidable sales tool in a hurried market that does not stop to properly think about requirements or goals) and fighting it to death. Which of course has contributed to simply expanding the conceptual black hole.

Final outcome? Take a deep dive at any Fortune 500 company out there and you will quickly find the following already in place for marketing data alone:

• Data visualization dashboards that expect normalized data – these are managed by IT or BI. Count on three or four different solutions at this level, as some business units/countries are more sensitive to specific SaaS sales teams
• Marketing dashboards with a built-in ETL (Extract, Transform, Load) layer for batch CSV uploads and unstructured data
• API-dependent dashboards, extremely agile at connecting multiple sources at scale but highly limited at defining cross-source metrics, breakdowns or exploratory capabilities
• Tactical dashboards for community managers and CRO professionals, directly plugged into one or more external data models and entirely dependent on them
• Wall-decorating dashboards that help us believe we are somehow marching forward…

So, no, you are definitely not alone in spending money on Tableau, PowerBI, Qlik, a social media dashboard, Google DataStudio, Adobe Workspace, an ETL-based platform (Sweetspot Intelligence or any other), and an API-based platform (Sweetspot Reports or any other)… SIMULTANEOUSLY.

But here’s the best part:

You will still find extensive use of Excel, PowerPoint and PDFs for internal reporting purposes. In fact, key goal tracking reports (of the sort that determine everyone’s bonus) are still being shared on spreadsheets, while conceptual (albeit data-driven) information is primarily delivered on PowerPoint.

In other words: for all the talk of automation and extensive abuse of dashboards, the very specific mission of Enterprise Reporting remains a manual process.

And I believe it boils down to two issues:

• A complete absence of basic contextual information (events, goals, benchmarks, forecasts)
• A discrimination of ideas and their representation in decision-support tools.

So, it is with all this in mind that, not turning our backs on the great value we believe we are already delivering for leading brands in different industries, we are setting forth a bolder framework for Enterprise Reporting in 2018.

CONTEXT, IDEAS, GOVERNANCE, VOICE.

Keeping it very simple, regardless of the chosen data model (ETL-based or not), every Sweetspot project will share two common sets of features as essential pillars of an effective Enterprise Reporting environment:

1. OBJECTIVE CONTEXT, referring to complementary information that provides basic actionability to regular KPI updates. This includes: events (campaigns, actions, changes, etc.), forecasts (ML-powered or not), benchmarks, and goals.
2. SUBJECTIVE CONTEXT, referring to either analyst-driven or automated recommendations, hypotheses, industry know-how, and team exchanges.

Their interplay results in even greater opportunities for storytelling and data verbalization, as well as an extreme focus on delivery (relevant updates coming to us wherever we happen to be), further steering away from an exclusive dependency on visualizations (great for analysis, not so good for strategic decision-support).

After all, we still believe that:

1. Self-service analytics has little to do with real analytics
2. Data democratization is an extremely dangerous game – everyone has an opinion on metrics they do not really understand
3. We all have enough on our plates already: Do we really have the time to log into yet another platform?

A first step in this direction has been the recent introduction of Natural Language Generation (automated headlines) and its combination with third party Natural Language Processing tools (conversational interfaces) across the entire Sweetspot offering.

I certainly do not expect you to replace or stop your many data integration, information delivery, campaign management, or visual discovery projects, as they all have a reason to exist, but I strongly believe Chief Digital Officer and CMO-led Enterprise Reporting will end up in far greater shape if they manage to look behind the curtains.

Setting ourselves free from the damn word will definitely be a long journey, and we will still need to make a living in the process.

Hence my sincere apologies for reverting to the compromise while happily pointing you to my favorite landing page ever: www.notanotherdashboard.com 🙂

May 2018 bring us less dashboards and more time to properly ponder what we really need.