What B2B SaaS Customers really want from Analytics and Dashboards

Dominic Denison
Dominic Denison
March 20, 2024

The article provides insights into the evolution of analytics in B2B SaaS applications, highlighting how it has transitioned from being an afterthought to a crucial component integrated into everyday processes. It emphasizes the shift from specialist-driven analytics to the expectation of everyday users to engage with and benefit from analytics seamlessly.

TABLE OF CONTENTS
A Person working with a System Embedding Analytics
Analytics
B2B
All
Product Management

A little history and the current reality for B2B SaaS applications

For decades analytics and dashboards were an “after-thought”, something management wanted (“Executive Dashboards”) to provide a top-down view on activities and processes in the company.

With the advent of ad hoc Business Intelligence (BI) tools, a new breed of analyst; data analysts and data scientists, were tasked with telling management teams stuff they didn’t know from flawed, or at least challenged, traditional operational (such as ERP, billing, reservation & rental, trading, retail and many other) systems as well as summary executive dashboards.  Those BI tools traditionally were separate and standalone from the operational systems they supported.

Thinking about your SaaS user personas

Elaborating briefly on data analysts and data scientists and their skill sets…Importantly, a small proportion of your users, depending on the B2B SaaS application, perhaps <5%, will be data analysts or scientists…

A data analyst is likely to understand data, how it’s structured and organised or modeled and probably how to query it.  A data scientist might have a statistical background or awareness on how to write predictive and other analytical algorithms, they could have a grasp of machine learning or artificial intelligence principles.  Both data analysts and data scientists would certainly be aware of which chart selections to use for a given data set (e.g. when to use a scatter, sankey, bar, line, pie or combo…?).

However, today’s SaaS users are everyday users, not typically data analysts and data scientists. They don’t just need analytics to please their management or to review and send end-of-period reports, they actually need analytics, operationally, day-in-day-out, to prioritise tasks, optimise everyday processes and get things done efficiently for themselves as well as others.

The challenge of data everywhere

The challenge is, particularly with the digitalization era, we’ve got really good at generating and collecting increasingly larger volumes of data (think eCommerce, EHR, automated trading, supply-chain & logistics, VoIP, IoT sensors for example - let’s not even start thinking about AI) and trawling through this has become essential for the growth and success, survival even, of any organisation.

Even public service organisations are expected to provide transparency through governments and municipalities as well as regulatory bodies, not only to know, but also to show how the tax-paying public (next year's electorate) is benefiting from their initiatives.  For more commercial entities it’s about differentiation and staying one-jump-ahead of competitors through maximising revenue growth opportunities whilst at the same time minimising costs to drive efficiency.

In short, data analytics is the weapon-of-choice for all organisations looking to drive improvement as well as thrive in competitive or stressed environments.

Did something change?

Yes.  Even though data volumes and variety (yes-yes, velocity, value and veracity too) have exploded, analytics in contemporary SaaS solutions are no longer the exclusive domain of specialists such as data analysts or data scientists.  Everyday business people expect and are expected to be on top of their relevant numbers, perhaps finding, even innovating or creating new indices/KPIs (LTV, NPS, RFM, OKR, EPS, ROI, BSC…) continually identifying problems/anomalies, understanding them and rectifying or optimising them.

This generation of analytics users are mostly not analysts at all.  They typically have no formal data or data science education, nor statistics and they won’t be trained on these either.  They’re less interested in analytics, if they were, they’d probably use a BI tool and call themselves analysts!  They’re far more interested in getting things done quickly, easily and efficiently than trawling through vast data sets to find obscure inconsistencies that need explaining.

Useful capabilities/resources for SaaS solution analytics

SaaS solution providers aren’t looking to “out-BI” BI-tools favoured by data analysts or data scientists, instead they look at how to make analytics an accessible, dynamic, intuitive, interactive, “don’t-even-think-about-it”, pervasive piece of most-if-not-all processes they are guiding users through in their SaaS solutions.

Younger generation users, in particular, are used to getting questions answered for themselves using on-line resources.  It’s normal then, that they expect, demand even, contextual, analytic feedback, at all times, as they navigate a SaaS solution or process.

The good news is that computer systems and analytic platforms are very well equipped to help these everyday users without the need for training or even experience or expertise.  How?

  • Hardware:  Computer systems have expanded exponentially in power and capacity and, better still, are available “on-demand” with seemingly infinite scale.
  • Elastic pricing:  With the gradual acceptance of cloud-based systems, it’s possible to dynamically right-size, that is dynamically requisition server resources, according to SaaS customer needs.
  • Analytic Technology:  In contrast to conventional SQL-based operational relational databases, analytic servers are optimised for analytic capabilities such as variances, comparisons, sophisticated weighting/scoring, business rules and management-by-exception thresholds.
  • Browser-based:  SaaS applications and their analytic platform brethren are available and accessible using similar user experience frameworks as well as extensible through APIs/PlugIns common to both.
  • White-labelling & Theming:  Obviously analytic content can be rendered in any part, or multiple parts of a web-page, typically via an iframe or a div.  Security and other parameters are passed to the analytic content to determine which analytic content may be rendered for a user.  All colours, fonts, themes and behaviour can and should be completely consistent with the host user experience.
  • Dynamic & interactive:  Sophisticated younger generation users are full of questions.  Giving these users a report as a flat-table for printing or export to a spreadsheet is not good enough.  They want “dynamic”, meaning that the data is correct and accurate to the current moment and interactive, meaning that the data can be navigated, if/as needed to understand the numbers-behind-the-numbers, the why-behind-the-what to validate or verify a credible and appropriate course-of-action.
  • Cross-platform:  It doesn’t matter much anymore whether you're accessing the SaaS, and its analytics through linux, Windows, MacOS on a desktop, tablet or phone nor even where, or in what format, the data resides.
  • Semantic Layer:  Building a semantic layer or a data model, simplifies the typical complexity of a physical data model for non-analyst users to understand and navigate.
  • Centralisation/consolidation:  By templating a semantic layer, this benefits from a build-once-use-many model which facilitates centralised maintenance in one or few places and means fewer “moving parts” to break.
  • Security & Governance:  A centralised or templated model also makes it easier to manage security and governance.  Security is normally (already) managed in the host SaaS application, this should not need to be duplicated and managed again in the analytics platform.  Typically this is a set-it-and-forget-it, on-the-fly, “handshake” between the host and the analytics.
  • Analytics-to-Host parameters:  We tend to think of the SaaS host “driving” analytic content, but if we want the analytics to drive action execution, particularly at scale, in the host, then we also need to have the analytic host pass parameters to the host to help us navigate to the right part of the host to do something (think reschedule, resend, reassign, reorder) and further pass parameters not only to execute one action but also multiple parameters for multiple actions.
  • Deployment Models:  Cloud models unfortunately do not remove all the complexity of customer deployment.  Happily the use of parameters, images, containers and orchestration tools such as Kubernetes, give DevOps specialists many choices to facilitate rapid and scalable deployment.
  • Integration with AI: AI is growing fast and promises to bring expertise into an organisation using it. See our AI / LLM thoughts for Data Analytics.

What does this mean for SaaS solution providers?

Analytics ≠ analysis

Where does this leave the intrepid SaaS solution provider?  Hopefully, we’re beginning to see that analytics is less about, or at least not just about, analysis and more about getting things done in your SaaS.

Another way of expressing this is “instead of answering questions…automating answers”.

If we’re thinking about “embedded analytics” simply as BI (for analysts) inserted into a SaaS product, that probably needs a rethink.  It wouldn’t be unreasonable to expect that at least some of your customers have that already.

User benefits from analytics

The reality is SaaS users benefit from analytics to:

  • Support them with dynamic, interactive, pervasive, immersed and ever-present (analytic) “good-to-know” feedback in their everyday in- SaaS processes, not just at summary (dashboard) level (analytics ≠ dashboards), but also at detail, perhaps record (a person, a product, a shelf, an order, an invoice, a customer…) level too.
  • Guide them through the shortest possible route(s) to the best possible outcomes
  • Help them execute the best possible outcomes in the quickest, easiest and most effective way possible and at scale.

Instead of asking only what do you want to analyse?  Let’s ask why you want to analyse and what improvements could you be looking to drive from analytics?

Host-to-analytics + analytics-to-host

Once we can answer these questions, we can link the relevant parameters and navigation up between the analytics and the host SaaS solution to automate, or at least semi-automate, using business rules and variable parameters, to drive the best possible improvements in the shortest possible time without relying on “propellor-heads” to drive decision-making processes.

Value for your SaaS

By.the-way, the SaaS solution provider (you) will be, concurrently developing priceless, differentiated intellectual property of value to investors as well as customers.

Analytics, as an integrated and embedded part of SaaS offerings, is an area of continual growth and innovation.  The complexity is in the back-end (data management), but the front-end (charts, filters, tables, maps…) is powerful, engaging “eye-candy” which catalyses and encourages customers to make decisions in favour of selecting your differentiated SaaS platform over another.

Would it be unrealistic to expect more new business (MRR/ARR), more quickly (shorter prospect->customer conversion lead-times), higher average deal value (differentiated premium value) and higher wallet-share expansions from existing customers for new premium (analytic) content?

You can find a more detailed analysis of Embedded Dashboards and their competitive advantage on this Forbes article.