Design and process improvements at a small software company

I joined TALtech in 2021 as the sole employee to help maintain and improve its software and platforms. Since 1985, their Windows products for automating instrument data collection have been mission critical for businesses ranging from small facilities to industry giants like Ford and Johnson & Johnson.

My role quickly expanded from Sales and Admin Assistant into an amorphous “everything-else” role juggling sales, support, design, and web dev. This position gave me a comprehensive view of their operations and revealed many opportunities for improvement.

Here I'll detail some of my work driving incremental design improvements and process optimizations within the constraints of a legacy environment and limited resources.

Key initiatives I led were:

1. Streamlining processes and documentation

2. Advocating for user-centered design

3. Tackling technical debt and system modernization

1. Streamlining docs and resources

Decades of unplanned growth led to scattered customer data, product specs, and procedures across multiple formats and locations.

The lack of a unified knowledge management approach made things hard to find:

• Customer purchase histories were split between an outdated Access database and email folders

• Product manuals existed in portions across network drives, often with conflicting information

• Internal processes, like order fulfillment steps, were inconsistently documented.

I saw chances to make incremental improvements and focused on the sales, admin and support workflows closest to my role. For instance, I digitized the employee guide as a searchable Notion wiki, cutting outdated information creating a central hub for newer how-tos, templates, and knowledge bases.

I also saw that our video tutorials needed a significant update. They were stuck in the 90s: low-resolutions, outdated Windows versions, and long runtimes. Referring customers to such old resources conflicted with our image as a provider of relevant, capable, and well-supported software.

I proposed and handled the production process to make more concise and polished videos. That involved writing scripts informed by my customer support conversations, recording voiceovers, creating animations, and editing final cuts.

These covered WinWedge setup, other software products, and general serial instrument connections.

The new vids…

1. Provided a clearer, more current resource for customers during the setup process

2. Reduced time spent on support calls and emails by offering a comprehensive visual guide

3. Acted as a compelling demo for prospective users, showcasing WinWedge's capabilities and flexibility

…But customers shouldn't need to lean so much on videos or guides to perform basic setup for our products. Ideally, better interface design should address any usability issues at their source.

2.1. Redesigning the Website: Updating TALtech's Online Presence

With a deeper understanding of TALtech's products / processes, I then turned to optimizing the user experience in the company marketing website and the WinWedge software interface.

The main marketing website had 90s-looking static, inconsistent layouts unsuited for smaller screens and navigation that changed unpredictably between pages. Its backend templating system, built on a deprecated Expression Engine CMS, was a maze of undocumented code fragments that made locating code for page changes extremely difficult and time-consuming.

I advocated that we switch to modern site building platforms like Framer or Webflow. These would empower TALtech a more approachable CMS with WYSIWYG editing. However, the president opted for WordPress due to its popularity and hired an external WordPress agency, with my role meant to assist throughout the process.

Before engaging the agency, the VP of Sales and I streamlined the sitemap so the agency would have a cleaner foundation to start from. We pruned over 500 pages and reorganized content to better suit customer needs based on support conversations.

But gaps in the process became apparent as the project unfolded. The contract lacked a dedicated wireframing phase, so I was concerned that the project was missing an way to ensure the agency was fully addressing our niche product features & user needs while drafting solutions.

To address this, I sketched some rough wireframes. These involved:

• Clarifying our desired structure of pages with bespoke headings, groupings, and section priorities based on our customer knowledge.

• Distinguished product use cases on the home page, as our tech-based software names had often sent customers calling to ask

• Defined hero illustrations that would clarify each product's compatible hardware and potential integration setups

Note: These rough sketches are internal communication tools. While they may appear unpolished, they were essential to getting everyone on the same page.

As implementation progressed, new challenges emerged:

1. The project scope left little room for brand strategy or mood boarding.

2. The agency's contracted designer appeared to be spread thin across multiple projects, leaning on an existing template he had built for another company. The resulting designs felt incongruously playful and consumer-oriented for our B2B, industrial, and research-leaning user base.

3. Communication lags with the designer sometimes stretched into weeks, eventually halting completely.

I stepped up to refine and complete the designs within the constraints of the agency's WordPress plugins and Elementor, a no-code design plugin within WordPress.

My improvements included:

• Developing a more professional color palette, balancing accessibility with visual appeal by adapting Tailwind's color system to maintain consistent saturation across lightness levels

• Enhancing layout consistency and responsiveness across the site

• Refining product illustrations for a more polished, trustworthy appearance

• Reframing technical descriptions to emphasize benefits over specifications and reorganizing feature comparisons for clarity

• Overhauling the support sections, including the creation of a step-by-step WinWedge Quick Start Guide and pruning of outdated articles

While WordPress is widely used, its reliance on plugins made some of my desired design improvements more limited than I hoped. Despite these constraints, the new site had meaningful outcomes: reduced support call volume, increased engagement with key pages, and smoother customer conversations now that we had better pages to reference.

Users could now more easily navigate our offerings and grasp the value of TALtech's software. The new design offered clearer wayfinding, better info, and a more polished look that better represented TALtech's professional capabilities.

2.2. Modernizing the flagship software

WinWedge streamlines data collection from measuring devices into destinations like Excel. Its flexibility is a core selling point, allowing users to customize dozens of settings for bespoke workflow automation. However, the complex interface can leave users struggling to navigate and comprehend effectively.

Previously, I had created a new walkthrough video and quick start article to clarify WinWedge's many windows and settings, including:

• Configuring “COM port settings” to decode device data

• Defining “input data record structures,” aka (1) “Fields” to separate data parts (ex. weight and unit), based on separators specific to the device; and (2) Keystrokes or Dynamic Data Exchange commands for sending to a destination application

But without those resources, the app itself offers no visual guidance to make setup easier to intuit. The high proportion of support requests for basic operations and quick fixes signaled the need for an interface overhaul.

Non-intuitive navigation and disjointed processes

The navigation relies entirely on the menu bar, forcing users through layers of fleeting menus and submenus. A menu bar can provide clear feature accessibility and consistency, but an exclusive reliance on it for all operations made it a hinderance. The setup process involves navigating multiple disconnected windows, each handling a different configuration aspect in isolation.

For example, the "Input Data Record Structure" window is separate from the "Port Analyze" window, despite their interdependence. The Analyze window is crucial for testing COM port settings and identifying working ASCII commands, but is depicted as a standalone tool rather than an integral part of the setup flow. We would tell customers to return to this window regularly while adjusting COM port settings, as incorrect settings would result in garbled output.

This disjointed approach causes users to lose context and continuity, leading to frustration, over-reliance on documentation, and ultimately, increased support calls.

Lack of written explanations and feature clarity

Compounding the navigation problems are insufficient explanations, inconsistent layouts, and unclear labels. The "Port Analyze" window exemplifies these problems:

• It presents various functions and labels data as "inputs" without clear explanations

• Ambiguous labels like "Input Data" and "Output" lack directional clarity and semantic symmetry—is data flow from WinWedge to the device, or the reverse?

• The absence of contextual help and actionable insights, combined with technical ASCII symbols, is daunting to non-technical users. "What are these for?"

This lack of clarity turned the software into a confusing toolbox, unnecessarily increasing the learning curve for new users.

I was asked to imagine a new approach to the WinWedge UI, and set out to design one that would guide users through the setup process logically—while preserving WinWedge's extensive options.

Implementation and code considerations for the redesign

I focused on optimizing the core interaction of configuring and managing communication settings. Many users only interact with WinWedge when hardware changes necessitate reconfiguration, so I aimed to make this process intuitive for non-IT specialists.

To reflect TALtech's emphasis on utility over artistry, I wanted to extend WinWedge's practical interface while aligning it with other system-level Windows apps. This approach would make the software more approachable and feel like a natural extension of the operating system. And as a software design enthusiast and a fan of Apple's Human Interface Guidelines, I valued this opportunity to better grasp Windows' design principles and patterns.

I dove into Microsoft's WinUI documentation and Figma design kit, internalizing their guidance on spacing, typography, component usage, layouts, and color. This not only improved the design but also aimed to simplify the eventual handoff to development by leveraging OS-provided components in WinUI 3.

I also considered the migration implications of the dated but widely-used Windows Dynamic Data Exchange system which WinWedge uses for non-keyboard integrations. WinUI 3 may not support DDE natively, but Windows often allows mixing backend frameworks. Could DDE work in a modern WinUI wrapper? But I caught myself overengineering prematurely. This project was only meant to be a guiding vision, so I refocused on creating a sleeker, more intuitive interface design.

Inspiration from similar utility apps

For inspiration on interaction flows, I looked to popular utility apps like Audio Hijack and Retrobatch, which feature friendly visual programming interfaces for data processing and file operation workflows. Their node-based workflow editors, where every feature, input, and output becomes a connectable block, sparked ideas for transforming transforming WinWedge into an intuitive *workflow builder* with a clear visual sequence of data processing steps.

Then I pivoted a bit. I realized that WinWedge's linear data flow didn't suit a freeform canvas, so then I looked to Apple Shortcuts' single-column approach. Apple Shortcuts' natural language writing, using fill-in-the-blank inputs within plain-English descriptions, aligned well with how our customers often described their needs in support calls: "I need to split the data using spaces into columns for the time, weight, and unit."

Single-window interface and object-oriented workflow editor

My initial redesign centers on a single-window interface with two areas

1. A primary, fixed canvas called the "data workflow"

2. A sidebar for accessing other settings and setup panels.

This layout emphasizes the users' main goal (building a device-to-PC data pipeline) while organizing the key setup steps and related settings into clear procedural sections on the side. It allows users to plan data workflows while testing device commands, minimizing the disorientation of the previous version.

A block-based, actions-oriented approach to building the data workflow offers several benefits:

• Visual clarity through nodes and connections, which make complex processes and relationships easier to grasp and debug.

• Intuitive editing via object-oriented, additive interactions

• Modularity and future extensibility: Each node is a self-contained action or function block, allowing for easy addition of new features and capabilities. As WinWedge evolves, TALtech could introduce modular features that users just drop into their workflows. Advanced blocks, monetized through add-ons or tiers, could enable flexible long-term growth guided by user needs.

• Progressive disclosure: Clicking on blocks reveals input flyouts with settings, reducing cognitive load and preventing overwhelming less knowledgeable users.

Some self-skepticism:

• If I were to continue iterating on this, I would reconsider making the block-based data workflow canvas completely rearrangeable. Some of the block actions actually ought to follow a specific sequence that potentially makes such flexibility unnecessary, even error-prone. Apple Shortcuts' pseudo-code approach to its workflow are undeniably powerful, but are object-oriented functions like "Get field" and "Save as variable" really useful in the context of data retrieval from serial devices? Dedicated user testing would be particularly valuable in these areas.

• I also considered including a refreshed version of the menu bar to ease the transition for longtime users, allowing interaction through both a new interface and the familiar menu hierarchy. But on closer inspection, I found that it risked muddling the new interface by distracting from the clear sequence of setup steps laid out in the sidebar navigation. I removed it for simplicity, pending user testing.

Improved copywriting, visual aids, and testing experience

I addressed smaller pain points with clearer headings and labels, action-oriented copywriting, and contextual explanations. Notable improvements include:

• Adding small icons to aid skimming and recognition.

• Renaming buttons to more clearly convey their functions; renaming “Port Analyzer” to "Port Inspector,” and making it accessible next to other settings.

• Relocating less-used features into subsections, providing some light context about their purpose.

• Other enhancements to the Port Analyzer include: (1) Clarifying the purpose of the "Command strings" section and providing guidance on using the ASCII chart for special characters; (2) Adding a "Common commands" button to surface pre-written ASCII commands based on frequent solutions from our support calls

Overall, the redesign aims to shift the user experiences from cumbersome to intuitive. By addressing key usability challenges and feedback, it has the potential to significantly improve customer setup times, reduce their reliance on support staff, and position the software as a more user-friendly solution for automated serial data collection. And by relying on native Windows system components and visual designs, the design sets up the development timeline to be potentially more cost-effective and efficient.

Leftovers and reflections

I never had to deal with so many unique constraints for a single "client" before—from untangling scattered documentation to reimagining core software interfaces. Without formal research tools, I gathered insights from customer interactions, support requests, and my own experiences creating user guides.

These design projects have dovetailed with a new focus on defining requirements for integrating e-commerce, license management, and customer data. This involves:

• Evaluating potential platforms like FastSpring and Stripe for e-commerce and licensing

• Outlining data migration strategies from our current fragmented systems; drafting project specs

• Working with the VP to prioritize features

• Assisting in the backend developer hiring process

• Designing a license manager that will allow customers to interact directly with our custom offline licensing system rather than having to contact support.

Working in the interplay between design, technology, and business constraints here has deepened my perspective on modernizing legacy environments. I'm eager to grow further in future roles where I can continue bridging the gap between user needs and technical possibilities.