Main content

Misinterpreting data comes with a price tag - Yellowfin's CEO explains why generic BI projects are off the table

Jon Reed Profile picture for user jreed April 23, 2020
I'm always up for a debate on how to derive value from data. In this case, Yellowfin's CEO stepped into the ring, with a clash of views on data literacy. He also shared how Yellowfin's customers are responding - and why generic BI is giving way to something better.

Data scientist in glasses and pink shirt through touchscreen display, with wood background © WAYHOME studio - shutterstock
(© WAYHOME studio - shutterstock)

Every pitch that dodges Gmail's spam filter and hits my inbox is created equal - but then it quickly goes in one of two directions.

Usually downhill, as you might guess. But: vive la exceptions. Yellowfin's PR team got it done with a practical pitch on our data conundrums:

  1. In this uncertain time, acting on the right data is crucial.
  2. But: data proliferation leads to unmanageable noise.
  3. How can businesses sort through this data influx effectively?
  4. And how do you avoid misinterpreting that data, and the poor decisions and miscommunication that inevitably follow from that?

Leadership requires data - and context

That last point was the final hook and sinker - so Yellowfin CEO Glen Rabie called in from his home zone in Melbourne, Australia to hammer it out. Rabie started here:

This situation has brought home to me: how do organizations, and leadership within organizations, motivate and lead their teams in times of crisis, or in difficult economic situations?

We might think a crisis would force concise and clear communication. Alas, that's hardly a given. Rabie:

I think what a lot of organizational leadership fails to do, is that when times are tough, we just tell people what to do - and we don't give them the context.

That's where data matters:

By the context, I mean we don't provide a narrative around it, we don't provide them with the data that underpins our decision making.

Rabie criticizes the "reactionary" media environment around COVID-19, and the questionable agendas behind the news stories. We certainly see that in the US right now, with conflicting messages from state and federal agencies. Rabie sees a similar problem in organizations:

You get this internal politics, which again, happens in organizations where you have one part of the organization wanting to react in a certain way, versus another part.

The media problem Rabie sees is not the airing of opinions, but the lack of a complete picture. He warns enterprises: we must do better

I think there are massive lessons to be learned here about how to use data, how to build narratives around it, and how to lead an organization through times of change like this.

CEOs are demanding better data

The one point Rabie left out that preoccupies me? Data literacy. If people don't have a hard-won core of critical thinking skills, It doesn't matter how good your data story is. In my experience, the necessary BS detector - one that comes without excessive cynicism, but includes the willingness to question your own assumptions - that's a rare model indeed. Rabie has a different twist on this:

I also see this situation actually driving data literacy. I think there's never been a more public conversation about data... More and more people are looking at data, thinking about it, talking about it.

On the one hand, what you're also seeing is people wanting to converge on what is really happening. I think the anxiety is driving that. That's making people question a little bit more, and wanting to go to a source - what is the reality? I think that that's a huge change from where we were historically.

Rabie believes CEOs everywhere are making a similar demand:

This will ripple through organizations. The conversation CEOs are having right now; I guarantee you they would be asking for - and demanding - more and more data from their entire organization.

When times are good; when your billings are going up, the CEO is simply not stressed. As long as the dollars are going up, I'm not really going to dig too hard. Now that I'm being asked by my board to fundamentally understand my risks and opportunities, I'm going to demand veracity of data - and I'll be demanding high quality. I'm going to demand where that data came from. I want to remove the bias.

Data bias comes in many flavors. CEOs must parse data from department heads with different agendas - sound familiar?

When I talk to my CFO, their view of the world is very different to my head of sales. Both of them have an agenda - it's clear. My job is to remove their agenda to get to the data, and to understand the data in its complete, clear form - so that we can act as a leadership team. Now we can make some real decisions.

What customers need - the data stakes are high

Given Yellowfin's "generate transformational value from your data" pledge, I would think their customers are pressing them on how to do this. Rabie:

Everyone that I'm talking to - our partners and our customers - they are experiencing a similar change within their organizations.

Misunderstand your data, and pay the price:

When you're forced to think about data in this way, when it's being driven through the organization, and people are asking more questions; people's working lives are at stake. They are far, far more about making sure it's accurate than they've ever been before.

People are fighting for their survival. Businesses are fighting for their survival. And yeah, that changes the dynamic.

And what do customers want from Yellowfin right now?

We fall into two types of customers. We have direct customers - enterprise customers who use us for data analytics. Then we have a very large partner base - people who build applications around us, or embed us into their products. For the partners that do the embedding and building, they are turning to us to say, "How do we turn this into an opportunity? How do we build things faster, and help our customers with turnkey solutions?"

Goodbye, generic BI

Elaborate/generic analytics projects are now completely off the table.

Historically, you could have gone into a customer and said, "Look, you need analytics. So let's do a twelve-month project." Right now, customers just want to turn stuff on to get the answers. And so I think there's going to be this massive shift from generic BI. and generic analytics.

So what's the alternative? Quicker ramp-ups with a vertical bent:

Here's a tool for someone coming in and selling you procurement analytics, IT analytics or healthcare analytics - all pre-configured. And yes, you may have to massage your data sources. But as an organization, I don't have twelve months to wait for someone to come up with a new idea. I just want to get going.

Rabie sees this as a fundamental shift in the analytics market. Prepackaged vertical analytics work especially well around common data sets, e.g. Salesforce, Marketo, or healthcare data that is governed to be collected in the same way. Seizing the opportunity to quickly benchmark across hospitals is one example.

Another key trend? Automated analytics. Yellowfin sees a surge in interest here.

We have an automated product called Yellowfin Signals, which does automated analysis - the demand for that has increased dramatically. Because again, I don't have the time to hope that my data analysts find something for me. I need to know what's happening at a really low level in my organization very quickly, to take advantage of those opportunities.

Demand questions are urgent:

If you've got an increase in growth of a certain product range, but everything else is dropping, I want to be able to understand why that's growing and to amplify that - rather than have that be hidden.

My take

An accelerated move towards lean/vertical/prepackaged analytics, with an emphasis on automated features? That's something customers have been gravitating towards for a while - I'm not surprised to hear the uptake on that now.

I found my clash with Rabie on data literacy refreshing - agreement makes for bland conclusions. Rabie's view is not just based on that increased desire for data. He also believes that modern BI tools should be inherently collaborative. That means users participate in building relevant "data stories," or narratives if you will. That helps users at all levels in the organization get a clear picture of what the data is telling them.

Yellowfin enables this type of collaboration via Yellowfin Stories. How that works is worth another detailed article at some point. But Rabie argues that just putting dashboards up and letting users come to their own conclusions isn't effective. Especially now - when a glance at a dashboard with red color all over it can conjure up a scary story, and lead to paralysis, or panicked decisions.

Now, more than ever, Rabie says, our data needs context. "We better provide context, and tell people what we're doing about it," he adds. He believes collaborative story-building with data can build trust, bringing data analysts and business analysts/domain experts together. That's better than firing off over-excited emails. Rabie:

It means that you build this level of transparency across the whole organization... This form of narrative helps people to organize their thoughts and ideas, versus shooting off an email.

Of course, the other part of data trust is data cleansing and governance - a topic I've covered recently but one that must be fused with this.

That leaves us with the question of data literacy. Several years ago, I grappled with this issue in On filter bubbles and AWS outages - does the enterprise have a fake news problem?

Bad enterprise decisions sink projects, organizations, and, eventually, careers; we can't afford to fall for fake news or poor data interpretation. Up until COVID-19, the stakes for getting the data right in our consumer decisions haven't been as high. Of course, that's changed now. Either way, organizations can do better - and now, we must. On that, I believe Rabie and I are in full agreement.

A grey colored placeholder image