The Ask

When your car repeatedly stalls out in traffic, it’s not always clear why. There’s gas in the tank and you’re standing on the accelerator, but the car won’t budge. The product org at Fitch Solutions was similarly stumped in 2019 as to their Web-based research platform, Fitch Connect. Why weren’t clients re-upping their subscriptions? All the warning lights on the “dash” pointed to design. So when I joined, it was time to pop the hood and take a look.

The Challenge

FitchConnect was a mess. It’s user experience was a tangled ball of cognitive overload and inscrutable workflows. To fix this, three interlocking challenges would need to be met in parallel. The first was to radically evolve the design culture. The second was to disabuse powerful stakeholders of the premise upon which the design of FitchConnect was conceived. Meeting these would enable us to meet the third challenge of untangling and improving the UX.

Why were big name clients like JP Morgan Chase not renewing their subscriptions to FitchConnect ?

What FitchConnect was supposed to do

Welcome to the world of financial news!

Did you know an entire shadow world of journalism exists in parallel to the mainstream news you get from outlets like CNN and The New York Times? If you’re like me, news is something you half listen to while eating your bowl of cereal in the morning.

But for high-finance types, news is the lifeblood of their multi-million dollar global investment strategy. So it has to be thorough. And not just Wall Street Journal thorough. We’re talking vetted, researched, corroborated, and highly, HIGHLY detailed thorough. Where can they go to get this kind of information?

Well, generally, Bloomberg. But also FitchConnect. At least this was the theory.

worth the money?

As you might have guessed, access to this kind of critical research was not cheap. Try $70K a year for a subscription!

But in 2019, FitchConnect (FC) was a research platform you needed additional research to use. For reasons I’ll get into, the non-designers who imagined this product seemed to have confused quantity with quality. Complexity with sophistication. And novelty with utility. All errors common to the novice.

And this junior design decision-making had gotten between users and the valuable information they had paid so much money to access. Believe it or not, the arcane, bifurcated wall of mysterious UI below was intended to facilitate one of the most common user experiences on the Web. So why does it look like it’s used to pilot the space shuttle?

No access to users

The biggest open secret in industry-specific enterprise software is that actually talking to actual users of the digital products being offered in this space is usually the last thing on anyone’s mind. In the highly secure, highly secretive world of high finance, this was doubly so.

When this door remains shut, product designers must seek out and identify those internal resources who either most closely resemble the target persona, or who most closely work with actual customers. At Fitch I found one of each and validated my design ideas with them.

Remote-only access to the product org

I myself, and the people who actually coded FitchConnect, worked out of office space in downtown Chicago. Everyone who decided what FC needed to do worked out of the Hearst Tower in New York City. In over 2 years, I never once met the Product Manager of FitchConnect in person.

I believe teams working remotely can be successful. I also believe it’s a little bananas to NOT get people who work together in the same physical room every now and then. Covid would change everything mid-project, but I was grateful for the time I’d had in person with the dev team at least.

Product Org preeminence

To appreciate the value of getting people who work together in the same room, one must first admit they are in fact working TOGETHER. This dynamic was more like: there’s the product org, and then there’s everyone who works FOR the product org. We in The Second City were in that second category.

Surprise, surprise, corporate politics exists. You won’t catch me complaining about it. My job was to understand it, and to the degree I could, work around it so everyone could be successful. Much of this, as we shall see, involved protecting the product org in New York from themselves.

As Product Management thought leader Marty Cagan has written, success in creating or innovating enterprise software depends on solving for three discrete, but interlocking risks. Often referred to as “The Holy Trinity,” they are as follows:

First, is the solution valuable? That is, can we sell it? Second, is the solution feasible? That is, can we build it with the resources we have, and in the time allotted?

And third, is the solution usable? That is, can a reasonably intelligent person sit down in front of this product, understand what it is, how it works, and, if they are the intended persona, start using it with little or no instruction?

Successfuly managing these three risks means making smart decisions about what to build and how to build it. Such decisions are said to be smart when each has been EQUALLY vetted for value, feasibility, and usability.

All such decisions should promote value, but none of the three considerations should otherwise predominate. We all want a perpetual motion machine, but if it can’t be built, it’s time to consider the next most valuable option.

So how does all this theory bear on an empowered product team?

In a word, ROLE. That is, who is chosen to tackle which group of our three kinds of risks. Because valuable, feasible, and usable aren’t just spheres of concern. They are spheres of expertise. For example, you wouldn’t want your lead developer regularly sitting down with clients all day to collect feedback or making presentations to leadership about strategy and roadmaps. That’s generally the Product Manager’s job.

Obviously, putting developers and technical people in charge of feasibility is not a tough call. They write the computer code that makes the product function, so they know best about what can be built and how long it’s likely to take.

A Product Manager’s role is not as popularly understood. But, as we’ve described, they handle the work of ensuring only valuable solutions get built.

Then, there are those who ensure usability — God help us, UX Designers such as myself. I argue UX occupies the least-understood third of the triad. Which can invite the kind of problems we were having at Fitch.

No model is perfect, but the SIZE and POSITION of each circle in the Venn diagram is significant.

The size is NOT a reflection of the number of people dedicated. Indeed most product teams have several developers, but only one Product Manager. Rather, it is a reflection of WEIGHT. The importance afforded each risk. The degree to which each is taken seriously, for any given decision, by other members of the team. And by the organization as a whole.

And the position of each circle is a reflection of that sector’s AUTONOMY relative to the other two.

We had the opposite problem at Fitch. This time PRODUCT predominated. The folks in Product Management at Fitch were whip-crack smart about the world of banking, investing, and especially credit ratings. They knew what this market would find valuable: intelligence! Investing is all about discovering opportunity and knowing something about it others don’t.

To Product, selling this intelligence was merely a matter of making it available on a Web-site. So the priority, above all, was getting that site built. Which to them meant just getting a group of developers to build it for them.

But build WHAT, exactly? Builders have questions. Managing feasibility entails more than just writing code. That’s why the valuable and feasible circles are supposed to overlap.

There was no UX Designer involved in the construction of FitchConnect as I found it in May of 2019. The design of its user experience was happening as a result of a defacto arrangement of shared responsibility between a few enterprising developers and the Product Manager of record — let’s call him “Sam.”

Product Management is a full-time job. Which meant Sam’s designs were always incomplete. The devs would get these designs and try their best to implement them in a way that made sense. This resulted in four years of anecdotal decisions and unholy compromises.

UX Design was happening as a result of a defacto arrangement of shared responsibility between the developers and the Product Manager

The Product Manager

“Sam,” was the Product Manager of FitchConnect (FC). He worked in tandem with a peer in New York who helped coordinate the efforts of some seven individual Product Owners. Each PO was responsible for one of the various specialized data or functions offered in FC.

Sam, in short, was a character. Which made him very engaging to work with — most of the time. Other times he was a man caught between a rock and a hard place.

As learned about the business end as any of his peers in New York, Sam was additionally very enthusiastic about the creative arts. He himself was an amateur photographer.

Sam was the person most directly responsible for the unfortunate state of his product’s user experience. With the help of some well-meaning developers, and a visual design contractor or two, he had podged FC’s UX together over the course of about four years. Mainly with the aim of impressing his superiors and indulging his creative impulses. Talking to actual users was not a thing at Fitch, so what else would he do?

The good news was that he welcomed change and enjoyed working with designers. But here were a few of the challenges of working with him :

His affection for the arts would often lead him to conflate UX Design with UI Design and perpetuate that misconception throughout the Product Org. Leading powerful people to assume that what the product needed was merely cosmetic. A re-skin. But FC’s UX problems ran much deeper.

To his credit, Sam surrendered a great deal of control over the design of his product. Which is not easy to do when you’ve been that involved for so long. I’ve described the difficulty that PMs can have making this leap in another case study.

But Sam did retain a veto power based less on evidence than on the capricious preferences of both his self and his superiors in New York. He embraced an interesting editorial voice HE HIMSELF likened to that of the notorious chief editor of Vogue Magazine, Anna Wintour. Leading to such arbitrary commands as “change that color, dahhling,” even if the color he wanted wasn’t remotely among those in the palette we’d defined.

The caution and deliberation with which I myself approached design impelled Sam to nickname me Martha Stewart, a media figure Sam presumably felt compared less favorably to the edgy and sophisticated Wintour. In Sam’s mind, my job was to impress him, the way a fashion designer would Wintour.

You can’t make this stuff up.

Why did Sam think of himself as an expert in design on par with Ms. Wintour? Because with the help of a visual design contractor and a few well-meaning developers, he had overseen the construction of a Web-site his bosses were sufficiently impressed with. And that clients were paying to access.

This fallacy runs deep in enterprise software. It’s easy to conflate correlation with causation when it seems to benefit you.

It typically goes like this: An unfortunate Product Manager, like Sam, is forced to do their own design work. Eventually they somehow manage to get a solution stood up. And because their clients NEED the content to do business, client users grudgingly push through bad usability. Ultimately, a resentful userbase is established who learn to cope with arduous training sessions, lengthy documentation, and protracted help desk tickets.

“Wow,” the Product Manager says, “look at all these users! Turns out I’m not such a bad designer after all!” But client users aren’t coming BECAUSE of the design, they’re coming in SPITE of it. The people who are SOLD enterprise software are rarely the same people who must USE it. And when those users decide the juice isn’t worth the squeeze, or find similar value elsewhere, they’ll stop using it. Their employer will notice the drop in usage and ask themselves why they’re paying for this expensive software subscription no one’s using.

And, as we’ve seen, this is exactly what was happening with FitchConnect.

Before I arrived, one man in Chicago understood the issues that were troubling FitchConnect. He had been part of empowered product teams in the past. Balanced teams that were not dominated by any one consideration. He saw the problems caused by foisting unfinished design artifacts and inadequate design processes on the Dev team. And of then leaving developers scrambling to account for gaps they’d had no hand in creating. That man was Mike Chesta, my boss at Fitch and the gentleman who hired me.

Mike and I became fast friends and remain so as I write years later. Luckily, we shared a lot of the same interests. Including an affection for bad action movies. Which led to shenanigans such as these (right) in October of 2019 for the office Halloween party.

Mike and I didn’t always agree, but we both understood the assignment: FitchConnect was unusable and it would go belly up if it kept bleeding subscriptions.

For two and a half years we made it our business to add balance to Fitch’s design culture, grow a legitimate UX Design practice, and pull FitchConnect out of the fire. I could not have accomplished anything at Fitch without Mike’s leadership.

I could not have accomplished anything at Fitch without Mike’s leadership

Was there a little bravado on display in the above photo? Assuredly. In 2019, Mike and I needed all the confidence we could muster. Together, we were about to do something a little risky: Imply to very powerful people in New York that perhaps they did NOT have all the answers. An endeavor that occasionally made scenes like this one (clip contains profanity and violence) from one of our favorite action movies, ‘Harley Davidson and the Marlboro Man,’ feel uncomfortably relatable.

Did WE have all the answers? No — far from it. Was there a villain in this movie? No, not at all. It’s hard to know what you don’t know. Fitch was full of truly brilliant and capable people, all doing their best with the information they had.

Were Mike and I offering a critical missing piece to the FitchConnect puzzle? I think that is undeniable. Could we have effected these changes by ourselves? Like in the movies? Guns blazing? Not in a million years.

Take a good look at this particular instance of the RLV below. This layout was used to present some of the most valuable content in the world. But what are you seeing on this page, exactly? Click here for a closeup.

If you guessed this was a list of the five most recent pieces of research relevant to the section of the site you’d navigated to, good guess.

But what about the rest of the page? The bottom half, that is. What are those pictures? Is this a Pinterest board of some kind? Forget about these for now.

Turning back to the research, tell me, how would you show MORE of it? That is, how would you expand the list? Think you have the answer? Okay, now tell me what you would expect to see next, once you’d made that choice?

Now stop and consider all this for just a second. Why are we THINKING so hard? We are not talking about exotic functionality here. We’re talking about simple pieces of written content. Research articles, mainly. Not very much different from the kind of stuff you’d see on your favorite news Web-site. Think about the last time you searched a Web-site for news articles related to a subject you were interested in. Keep that picture in your mind as we review three key problems with this view:

1) How do I load more articles? First, why only five articles to start? Second, assuming the user is curious enough to locate it, why is the link to load more at the TOP of the list, instead of the bottom? Third, clicking that link takes the user OFF this page entirely and sends them to the advanced search experience I described earlier. Which we’ve established is confusing enough as it is without being sent there unexpectedly. Suppose you turned a faucet expecting more water and instead a trap door opend beneath you and dumped you in a river. Good experience?

2) Where do I look? Just like the advanced search experience we visited earlier, this page seems to have two discrete sections, equal in prominence, one on top of the other. There is no visual hierarchy. Turns out the Pinterest-like photo tiles in the bottom half are just links to related material the system thinks the user might find interesting given the page they’re on. But would a user make this connection given the layout? A quick trip to any news Web-site shows related and/or ancillary material exposed to the SIDE of the primary, expected content, in a manner that reduces its prominence.

3) Why a table? When you look for articles on a news Web-site are they typically presented in a GRID like you’re there to balance your budget? No, of course not. Articles by their nature are intended to be READ. Or at least scanned. And when exposing lists of articles arranged by title, it follows you’ll want to do so in such a way that promotes legibility.

Now put yourself in the shoes of a FitchConnect user. Perhaps you are an investment manager. You need information.

For $70K a year you should be seeing an easy-to-enter gateway to all the research you need. You understand that Fitch offers the kind of research investors use to inform million, sometimes BILLION-dollar decisions, routinely affecting global markets, foreign governments, and regional economies.

So why is this wonky page making all that so difficult?

How on Earth had this bizarre design come to exist?

How on Earth had this bizarre design come to exist?

So it follows that someone in the Wall Street milieu, when tasked with imagining a new information-gathering tool for investors, would try and replicate the Bloomberg experience. And in so doing, hopefully, replicate that success.

Who was most deeply immersed in that milieu? I learned early that it was most likely Sam’s boss “Ethan,” whose affection for Bloomberg Terminals was made known to me.

Sam used to wryly refer to Ethan as an “influencer,” a sharp play on words. Ethan didn’t have a podcast. Rather, he was a powerful person at Fitch with strong views and an equally strong willingness to enforce conformity with these views. So it wasn’t too hard to figure out what had happened: At Ethan’s urging, either spoken or unspoken, Sam had used the Bloomberg Terminal as inspiration for the design of FitchConnect.

There was just one problem with that theory, obviously. These Bloomberg Terminals were not designed to provide a Web experience intended to expose research. But to instead expose raw data feeds from Bloomberg for the purpose of trading stocks. We couldn’t be 100% sure, because, as I’ve mentioned, actually talking to our users was verboten, but Mike and I proceeded on the assumption that Web-sites are inherently a different thing. And that not even most of our users wore suspenders

We weren’t just guessing, either. I checked the Web-sites for Moody’s, and S&P, the other two of the Big Three business credit rating agencies. Neither were offering research this way. Nor was Thomson Reuters. Or even Bloomberg itself! Bloomberg News does not present their articles at all the way their terminals present data.

Remember what I’d said about bravado and taking risks? Mike Chesta and I were challenging the design impulses of Sam’s boss, Ethan — one of the most powerful and respected people at Fitch Solutions. We were suggesting the entire theory that had driven his and Sam’s design strategy was without merit.

To Ethan’s credit, he kept an open mind and said he would absolutely yield to evidence.

And soon he would have it.

One was a really gracious gentleman who at that time was about to be promoted to the Director of Customer Success. He checked my box as being someone close enough to our users to understand what they were looking for.

The other was an extremely impressive woman who still serves as a Managing Director at Fitch Solutions. She checked my box as a person who regularly uses products like FitchConnect.

Both were generous enough with their time to sit for a usability test I’d written using a clickable prototype. Both navigated the proposed design as expected and each lent their unofficial stamp of approval to the approach overall.

In the winter of 2019, Mike and I waited with baited breath while the CEO himself reviewed the prototype and considered the evidence.

He gave us the green light.

A month or two after the new RLV’s implementation, some additional validation would land in our laps. Ironically, it wasn’t scientifically significant evidence, but the kind that moves mountains at Fitch — and probably the wider financial world in general: An unsolicited testimonial from someone of influence. About the RLV redesign, he remarked “this is the best new thing I’ve ever seen go into FitchConnect.”

Try entering the term ‘Brazil into the global search field.
Or view more of my prototypes.

Topics page - Anticipating a more Mobile use-case

Solutions to two of my long-term goals for the redesign of FitchConnect are on display in this page.

1. First, to radically reduce the amount of verbiage a user encountered per page until they got to the piece of research they were looking for. To otherwise guide them with imagery and logical page flow.

2. And second, prepare for a time when we could consider offering a more responsive, phone-friendly version of the site in which we would detect their device and beautifully serve WHAT CONTENT WE COULD in that context. So on a phone, the tiles on this page would stack VERTICALLY instead of cascade into rows, as we see below. So much of what we offered was just written articles. There was no reason we couldn’t enable diligent users to check the news from Fitch as they rode the train into work every morning.

Certainly the most complex of the views Sam and I tackled, “Financials” were a deep dive on the true worth of a given company or country. The trick to redesigning this view was understanding what it all meant and how each piece connected to the rest. This meant a lot of stakeholder interview time with Sam and others.

Would this have all gone faster if Fitch’s UX Designer, myself, was also an expert in Wall Street banking, accounting, and taxes? Yes, but generally, those two skillsets are pretty mutually exclusive. I raise this point by way of mentioning some of the talk I’d guessed was happening in New York. Something was brewing at corporate.

Or view all of my online wireframe samples.
Or learn more about wireframes and how they fit in to the modern UX Design process.

Sometime in the spring, perhaps, of 2021, I got a very unusual Zoom invite. It was from Mike’s boss, the CTO. Nice enough guy but we rarely talked. He had clinically the lowest levels of interest in UX Design I think I’ve ever measured in the wild. A year earlier I had requested a designated designer space with a door. He literally offered us the cloak room. Without saying as much, he was clearly relaying a request from Product to feel out my interest in the idea of reporting to Sam ( or perhaps someone else in Product ), instead of Mike.

It’s nice to be wanted. And if I thought switching would have helped anything at Fitch, it would have been a much harder decision. But it wouldn’t have helped. It would have just made me beholden to the other suboptimal sphere of the valuable, feasible, usable triad.

I told the CTO what I’ve said here: that UX is its own thing. That I’d love to report to a director of UX, if he wanted to somehow convey that sentiment.

Till then, Mike understood better there than anyone what I should be doing and how I should be doing it. And more, as I’ve said, I just plain liked Mike. We had a very constructive rapport. I’d never even met Sam in person!

I pride myself on being able to make design work in almost any context. Whether I report to UX, Product, or Engineering. I’ve done all three. And if they’d switched me without asking I would have made the best of it.

But they asked me what I THOUGHT. And I told them.

Evolving the Design Culture at Fitch Solutions

A redesign cannot happen in a vacuum.

And as I’ve said, this was not just Mike and I doing this all by ourselves.

Every enterprise software shop has a “design culture.” Sometimes the culture is based on ignoring design as a serious topic. This is when technical people and business people just kind of make anecdotal decisions about how their software should present to users and otherwise struggle through. That is an example of a design culture. A bad one. A dysfunctional one. But certainly not an uncommon one.

When I started at Fitch, as we’ve seen, things weren’t quite THAT bad. Leadership, and especially the Product Manager, appreciated the significance of design as a concern. As did a few of the developers. But mostly this amounted to lip-service. So what outcome was Fitch seeing from this very limited design culture?

Well, design was happening as a result of harried Product Owners making an endless stream of anecdotal decisions and then working with Developers to somehow force it through implementation. The outcome, as we’ve seen, was an ugly, fragmented, and often inscrutable User Experience that was costing the company subscriptions.

demystifying the mystification mindset

If you’ve ever dealt with an auto mechanic, you have dealt with someone who makes their living off specialized knowledge. Are they telling you the whole truth? Or are they keeping you in the dark to leverage you for more money? At your job, it’s less about money than prestige, politics and control. Often, the temptation to withhold information or otherwise “mystify” others as to your area of expertise can be hard to resist. This is true of Developers. And especially of high-finance types. Both had designed FitchConnect as I found it in 2019.

The result was poor crossfunctional communication, and a book of riddles masquerading as software. To break this cycle I led by example, freely explaining design theory, processes, and methodologies. “Shared understanding,” as Jeff Gothelf evangelizes, would be the new coin of the realm. There would be no more withholding of information on my watch — from eachother, or our users.

Design is a thing

In software, as we’ve seen, Design is a discrete concern, distinct from matters strictly of value to the customer or of implementation. This is to say that Product Managers are at their best when discovering what capabilities will be valuable to the user. And Developers are at their best when they are building the features or services that will afford said capabilities. Both get themselves in trouble when they additionally try to DESIGN these features or services themselves. This is gospel for thought leaders like Marty Cagan and Jeff Patton.

So I socialized the idea that Product Managers should push back on design they feel doesn’t facilitate desired capabilities. And that Developerss should push back on design they think will take too long to implement. But that they should both expect a conversation with UX when either presented objections of any other kind. This shift steadily moved us toward a far “holier” dynamic ( right ).

We don’t need another hero

Despite Jeff Gothelf’s best efforts, the concepts of “BDUF” ( Big Design Up-Front ) and “Design Heroes” persist like a plague in enterprise software and in 2019, Fitch Solutions was no exception. When I started, the notion that ONE skilled Designer could, in just a few months, meaningfully transform Fitch’s massive business application by aiming the right number of Photoshop comps at it, was very much the reality.

But by starting small and focusing on actual changes to actual production code, Mike and I were able to operationalize the iterative, collaborative process Gothelf describes in his industry-shifting book, Lean UX. And in so doing, we were able to build a rythm over time and maintain momentum .

When I left, the value of collaboration and iteration was fully embraced at Fitch.

Design Strata

When I started, exactly none of my stakeholders understood the distinction between Visual Design and Interaction Design ( except for Mike ). It’s not their fault — without a Designer around, why would they?

Product Owners did not understand why they couldn’t just attach a couple Balsamiqs to their tickets and trust Engineering to fill in the gaps. Surprise, surprise, developers don’t code anything you don’t explicitly tell them to.

Changing this mindset involved a gentle campaign of evangelism, racking up wins, and getting buy-in from Dev. Because what I knew from long experience was that Devs greatly prefer detailed wireframes over imagining use-cases.
After a while, everyone was using the term ‘Interaction Design’ appropriately. And as Marty Cagan has advised, it’s made for far better expectations — and outcomes.

Less is More

In enterprise software, not every page can be a dashboard. Once your user has selected a path, you have to stop showing off and let her get some actual work done. She can’t do this with a lot of ancillary junk in her face on every single page. The effect becomes that of a pinball machine. This kind of wanton overload was on full display throughout FitchConnect when I arrived in 2019.

Real designers understand that the mark of excellence in design is as much about what you take away as what you add. And that whatever is left must be arranged according to a hierarchy of visual prominence.

Referring to overwhelming legacy views as ‘inelegant’ earned me a few chuckles from those in both business and engineering when I started. But it wasn’t long before they were using that word regularly themselves.

How seriously? Leadership filled the job by hiring away S&P Global’s UX Director — let’s call him “Nathan.” Should they have promoted Mike or even myself to this directorship instead? Arguably. But, ironically, Nathan’s appointment to the position was a far bigger acknowledgment of our success. As an S&P guy, Nathan was one of their own. A fellow master of the universe. Putting him on the job showed their newfound reverance for UX more than any other move they could have made. The stakes, they now understood, were far higher than they’d presumed.

It is difficult to overstate the significance of this organizational shift. In my two and a half years at Fitch Solutions, leadership had gone from kidding themselves that smart people with no design experience could just make up UX Design as they went along, to formally establishing UX as a discrete, independent, and permanent office of concern. Mike and I had backed up our evangelism with results. In 2021, usage was already up — way up. We had successfully proven the ROI of serious User Experience Design.