CASE STUDY 01 · DoD / DEFENSE TECHNICAL INFORMATION CENTER
Everyone knew where the gated content started. Nobody had told the users.
A Defense Department research service had not grown its user base in years. Leadership believed the market was simply small. An audit of the experience found something else: the service's most valuable content sat behind registration, and nothing told visitors it was there. I proved the point with a single email, fixed the content and interface behind it, built lifecycle campaigns on a segmented user list, and instrumented everything so we could tell what worked.
- Role
- Marketing and communications lead, 2016 to 2019
- Scope
- User research, user experience, content strategy, service design, lifecycle marketing, measurement
- Tools
- Email marketing · Web analytics · UTM campaign tagging · SharePoint
- 59%Registered-user growth in one year, 33,159 to 52,589
- 2,303Lapsed users returned, a 21.3% response rate
- 4xThe previous record yearly increase, exceeded
- 5,950Unique users in journal launch week, up 3,382
02 / In short
The project at a glance
- Organization
- U.S. Department of Defense, Defense Technical Information Center
- Role
- Marketing and communications lead
- Period
- 2016 to 2019
- Disciplines
- User research, user experience, content strategy, service design, lifecycle marketing, measurement
- Primary audiences
- Defense scientists, engineers, program managers, contractors, and academic partners
- Outcomes
- Registered users grew from 33,159 to 52,589 in one year, 2,303 lapsed users returned at a 21.3% response rate, and a tagline and access-first page structure still in use nearly a decade later
03 / The situation
Flat for years, and everyone had an explanation
DTIC stores and disseminates the Defense Department's extensive collection of controlled-unclassified technical reports and research projects. Scientists, engineers, program managers, and contractors use it to access, share, and build on past and current knowledge, innovate faster, and reduce duplication of effort.
Registration had been flat for years. My supervisor later described it in my performance record as "several years of near zero gains."
Leadership had an explanation. R&E Gateway, the flagship product, served a niche audience, and a small pool means flat growth. So the team tracked usage, how much existing account holders used the service, and watched neither how many new people arrived nor how many quietly left.
I did not think the ceiling was real.
04 / What I found
The gate was invisible to everyone who could fix it
DTIC's most comprehensive material sits behind registration, and nothing told visitors that. A first-time visitor saw the public layer, assumed that was all there was, decided it was not worth much, and left.
People told us directly.
DTIC User Council members, and researchers my team met at conferences, kept saying the same thing about the R&E Gateway landing page. It looked dated. It was not appealing. Nobody had framed it as a business problem, so nobody had treated it as one.
Then I audited the experience myself.
I went through every page, product, and piece of marketing collateral we had, including R&E Gateway, DoDTechipedia, Search, and the Information Analysis Center content. I put myself in the user's shoes and moved through it the way a first-time visitor would, not the way an insider does.
The same failures repeated everywhere.
Navigation took insider knowledge. The writing was dense and full of jargon, written by specialists for specialists. No product said plainly what it was worth, who it was for, or how someone would use it. Access requirements were unclear. And nothing showed users where free content ended and gated content began.
Leadership already knew about the gap.
It was so obvious to them that they had stopped seeing it as a problem, and they assumed nothing could be done. That is the oldest trap in this work. The people closest to a service cannot see it the way a newcomer does, so what confuses every visitor becomes invisible to everyone who could fix it.
Registration was not flat because the audience was small. Registration was flat because we had never told anyone what they were missing.
05 / What I did
Prove it cheaply, then build behind the proof
1. I tested the cheapest version of the idea first
I did not ask for a redesign, a budget, or a platform. I wrote an email.
It reminded users that signing in unlocked substantially more than the public side, and made the gap concrete by naming the size of the full research collection next to the public one. It carried a link for people who had forgotten their username, a link to register for anyone without an account, and a request to forward it to colleagues.
Leadership approved it. Registered users started going up.
That email cost nothing and answered the only question that mattered: was the gated content gap actually suppressing registration, or was I wrong. Everything after it was easier to justify.
2. In parallel, I fixed the experience the message pointed at
An email can tell people what they are missing. It cannot make the service easy to use once they arrive.
I rebuilt how DTIC explained itself. Every product had to answer four questions in plain language before anything else: who this is for, what it does for you, what it takes to get in, and what you unlock when you register.
The redesign ran as a long sequence of small phases rather than one launch. We changed a little, put it in front of users, listened, and changed again. User Council members reviewed every beta before it shipped. The fully redesigned page launched in 2017 and moved the Gateway to a modern, user-centered interface, and we kept revising it each year after. I rebuilt the collateral on the same model, so a postcard handed to someone at a conference made the same promise the website did.
The tagline came out of that work: Preserving knowledge | Connecting people | Inspiring innovation. DTIC still uses it several redesigns later, along with some of the content I developed. Visitors no longer have to guess which side of the access line they are on.
What I chose not to do: buy more advertising. More traffic into a funnel that could not convert would have burned budget and proved leadership's point about the ceiling.
3. I stopped treating every user the same
The email worked because it reached people who already had some relationship with DTIC, which told me the list was an asset nobody was using well.
I segmented it: power users, occasional users, dormant users, first-time and last-time users, newsletter subscribers, DTIC-only users, users of both DTIC and the Information Analysis Centers, and IAC users with no DTIC account at all. That last group was a standing opportunity hiding in plain sight.
Reactivation. I built a process to identify account holders drifting toward dormancy, then brought them back with the value proposition we had just made clear. In its first eight months, the campaign returned 2,303 lapsed users at a 21.3% response rate.
Renewal, then automation. We began tracking accounts at 30, 60, and 90 days from expiration and reminding users to renew before they lost access. Then I worked with our developers and the metrics team to build that tracking into the platform. Once a user is flagged, the system generates the monthly list and sends the reminder on its own. Retention stopped depending on anyone remembering to look.
Leadership had been watching usage. Nobody had been watching attrition.
4. I took it outside, selectively
Leadership saw conferences as expensive outreach with little return. They were right about the events we had been attending and wrong about conferences as a channel.
I set selection criteria: attendee demographics, event size, research interests, and cost, matched against who we needed to reach. Then I stopped treating events as booth duty. Each became a campaign, with advertising and email before, engagement during, and measurement after, tracking registrations back to the event that produced them. The Pacific Operations Science and Technology Conference and the Science and Engineering Technology Conference both met the criteria and contributed to growth year after year.
I applied the same logic to partnerships. The IAC Customer Upsell Campaign put DTIC's full offering in front of people already using one part of it, and I built cross-sell campaigns with DLA, DTRA, DISA, WHS, and DoDEA, trading social posts, newsletter mentions, and ad placements to reach audiences we could not reach alone.
Throughout: instrument everything
I introduced UTM parameters across everything my team produced, so every ad, email, banner, and link reported back on what it did. For the launch of the Journal of DoD Research & Engineering, I built the measurement framework myself: 27 click tracking tags capturing 57 data points across every channel. The metrics team adopted it as the foundation for management's weekly reporting.
Launch week drew 5,950 unique users, up 3,382 over the previous week.
06 / Results
Two years of growth at a service that had not grown in years.
The cleanest read is the reactivation campaign, because it is a closed loop. A defined list, a defined message, a measured response: 2,303 users returned at a 21.3% response rate.
Several efforts ran at once, so no single tactic owns the 59%. What the record shows is a service that had not grown in years growing two years running, once someone treated flat registration as a question worth asking.
Registered-user growth in one year
From 33,159 to 52,589, documented by my supervisor as more than four times the previous record yearly increase
Growth the year before
The first full year of the content and brand work, after years of nearly no growth at all
Lapsed users reactivated
A 21.3% response rate in the campaign's first eight months
Unique users in journal launch week
Up 3,382 over the previous week, attributable because the tracking framework existed
Measurement framework built from scratch
Fifty-seven data points across every channel, adopted by the metrics team for management's weekly reporting
- • The work stuck: nearly a decade later, through multiple redesigns and a department renaming, DTIC's landing page still carries the tagline I wrote in its logo lockup
- • It still leads with the headline I wrote, and still opens by telling visitors what credentials they need, what sits behind the gate, and where to register
- • The specific campaigns ended. The structure did not
07 / What I took from it
What I learned
Lesson 01
Being close to a service makes you the worst judge of it.
Leadership was not careless. They were expert. They knew where the gated content began, so they could not feel what it was like not to know. Every team I have worked with since has a version of this blind spot, and the fastest way to find it is to walk the experience yourself as a stranger would.
Lesson 02
Without data, marketing cannot move numbers.
The team cannot tell what to prioritize or what to cut. Everyone defends a favorite tactic and nothing improves. The UTM work was not a reporting chore. It was what let me stop doing the things that did not work.
Lesson 03
The cheapest test beats the best argument.
I could have spent months making the case for a redesign. One email settled it in a week and paid for everything that came after.
08 / What I would do differently
Capture the baseline before changing anything
I kept careful numbers on this work and almost no visual record of what the experience looked like before I touched it. I recovered the images on this page from web archives years later, which worked, but only by luck. Screenshots cost nothing at the time and are close to impossible to recreate afterward.
09 / Where it stands now
The structure held up better than the content inside it
Looking at the page today, the structure I built has held up better than the content inside it. The first sentence a new visitor reads now carries five acronyms before it says anything useful, which quietly tells anyone outside the community that this service is not for them. The one number that made the original campaign work, the size of the collection, is gone from the page. Five buttons compete at equal weight with no clear first step, and one of them leads to a search that is currently offline.
None of that is a failure of the redesign. It is what happens to every interface without someone whose job is to keep asking what a stranger sees. That role is the work. It does not end at launch.