The Problem

In 2018, Discover Durham transitioned their photo library from a series of network drive folders to Photoshelter — a fully cloud-based digital asset management system. This change brought a number of benefits: it was easy to browse and find images, search for them based on tags, and access the library from anywhere.

However, after this initial honeymoon period, several issues arose. Throughout the organization, users would have difficulty finding images in the DAM and would ask the creative team for help — after which we’d be able to find the photos easily. These requests would happen at least weekly — meaning we were eventually searching the DAM on these users’ behalf, which erased the value of the tool. Even when users could locate images they needed, it was difficult for our users to discern if the files were cleared with the appropriate licensing for their intended use. Once again, they needed to consult the creative team on usage rights, which added weekly back-and-forth. These cumulative inquiries were adding work to our plate and quickly invalidating the self-service model for the DAM.

In light of these accumulating issues, in 2020 I started a project to identify the systemic breakdowns among our users and design solutions that could close the gap.

My Role, the Team, and the Scope

I was the primary researcher, UX designer, and interim asset manager in this review and overhaul of the digital asset management system. I collaborated with the two designers (primary asset managers) on the creative team for their institutional knowledge and insights. I consulted with the other internal users in the organization for usability reviews and their downstream stakeholder insights. The revamped metadata schema built upon a foundation left by my predecessor.

There was zero budget and no development resources allocated for the usability revamp. Any solutions we identified needed to be accomplished within existing platforms and tools we already had or could acquire for free / inexpensively. In the end, I purchased only one Lightroom plugin to aid in building out one of the design solutions.

Ideating and testing new solutions

Based on the user interviews and user flows, I was able to chat with one of our asset managers and brainstorm different approaches we could take to solve the three broken processes (login, search, and rights confirmation). We came up with 12 different solutions, and after evaluating each one, I narrowed it down to 6 we could reasonably implement and would yield progress.

The rights confirmation process was the most broken and ambiguous part of most of the user flows. To address this, I researched Photoshelter’s custom metadata capabilities. This functionality could add a new metadata field to each photo listing and add a filterable dropdown to Photoshelter’s search interface. I experimented by uploading a small batch of images to the DAM and applying custom metadata to each that explicitly stated if they were “available for media & partner reuse” or “free use with photo credit.” I then had users perform a test exercise where they needed to search for these images and determine which ones they could download and share with journalists. Users were overwhelmingly able to identify the purpose of the field and use it to filter out irrelevant images — with only those with the appropriate usage permissions remaining.

I also tested out adding category keywords and more robust captions to our images. With these additions, searching users were more likely to discover photos they needed in their search results due to the additional text Photoshelter’s search was able to crawl.

The final broken process to address was logging in: many users found it too difficult to remember or look up the shared login info, so they just accessed the photos available on the public-facing side of the DAM. In response, I created individual accounts for each user and controlled their permissions using user groups in Photoshelter. This allowed each user to easily manage and remember their own login and password info, while the asset managers could control access at a group level without having to micromanage the permissions for 20+ different accounts.

As I tested out different approaches with our end-users, I soon realized our asset managers formed a second set of users that had their own needs to be met. We had to develop a corresponding asset manager workflow to supply and maintain the metadata for each solution we built out for the end-users.

This management component became quickly apparent when I realized there were legal-related metadata that we needed to include with some of our assets to meet the expectations of our operations team (a group of internal stakeholders) and reduce risk to the organization. Upon introducing a new type of metadata that we needed to support, I was able to see firsthand how inefficient our existing metadata practices were.

We used Adobe Lightroom Classic to maintain a catalog of all of our photos. When we applied metadata to new images, the asset managers had to toggle through three different views of the metadata panel to access all of the fields they needed to write information. Many of the metadata panel fields were extraneous to our needs or redundant with fields in another view. These hidden views and visual noise resulted in a heavy cognitive load for the asset managers and made metadata management an error-prone process.

To overcome these issues, I implemented two tools within our Lightroom catalog. First, I purchased a plugin to create a new default metadata view that contained all of the fields that our asset managers needed to access (while filtering out all of the fields they didn’t need to see). Second, I made a suite of Lightroom metadata presets that could automatically write metadata to our photos upon importing them into the catalog. So much of the metadata we needed to write was uniform or varied slightly based on the photographer. By making a preset for each photographer, we could write a majority of the metadata to the photos in bulk without customizing each piece of metadata.

These two tools freed up the asset managers to focus on the minority of data that could not be automatically written but would provide the most value to our users and internal stakeholders: titles, captions, category keywords, and legal paperwork. After adding the custom metadata panel and the presets to the workflow, we were able to reduce the asset management process from 13 steps to 11 steps, and we reduced the amount of metadata that had to be manually written for a photo from 16 fields to 7 fields.

Since rolling out the new metadata schema and resources, the creative team has seen a massive decline in photography inquiries from internal users. Prior to the revamp, users asked almost weekly about which photos had what usage permissions — now they ask those kinds of questions once every couple of months in very specific use cases.

We used to also have monthly false positives of a user believing we didn’t have an asset in our DAM because they couldn’t find it through searching. Similarly, those inquiries are all but extinguished. Now, the majority of requests that the creative team receives are for photography that is genuinely missing from the library, allowing the team to focus on prioritizing new image acquisitions that will more fully meet users’ strategic needs.

The new workflows have also yielded dividends on the asset management side of the equation. With the new workflow, the creative team imported, cataloged, and uploaded nearly 9,000 photos to the DAM in the 18 months after the initial launch (a more than 350% increase in available photos). This massive swell in the library was made possible by the streamlined workflow.