The Velocity Constraint
As our AI overlords becomes stronger and stronger, the age old bottleneck of testing becomes centre stage again. Explore the data-driven reality of the modern SDLC bottleneck.
The Capacity Crunch
Despite the gripes we may have with it, the modern developer is no longer a code-author primarily. Data from the GitLab 2024 DevSecOps Report reveals a shocking imbalance: developers now spend less than a quarter of their week writing new features, with the remainder lost to validation and fixes.
How Developers Spend Their Time
Source: GitLab DevSecOps Survey 2024. Only 21% of a developer's time is spent on new code, as testing and debugging consume the lion's share of cycles.
Sprint Velocity Loss
Teams suffering from QA spillover deliver 35% fewer story points per quarter. Manual testing isn't just slow; it directly neutralizes one-third of engineering capacity.
The Bottleneck King
of Agile teams explicitly identify testing as their primary sprint bottleneck, according to the State of Agile Report 2025.
The Financial Fault Line
If only we all had crystal balls when planning features. The worst time to catch a bug is in prodction, the next worst time is right now (I'm sure that's how the saying goes). Waiting until the end of the cycle to test isn't just inefficient — it's financially reckless. The cost of remediating defects escalates exponentially as code moves from design to production.
The Multiplier Effect: Cost to Fix Defects
Source: IBM Systems Sciences Institute. Fixing a production bug costs up to 100x more than finding it during design. Manual QA typically finds bugs at the 15x threshold.
US Economy Loss (CISQ)
The Maintenance Treadmill
I know what you're thinking, "who even does manual testing anymore?" "All I need is a few scripts and I can automate the whole thing!" Well do I have a surprise for you! Even "automated" teams struggle. Scripted automation hits a structural ceiling of roughly 25% coverage, forcing engineers into a cycle of manual maintenance and "flaky" test triage.
The 25% Ceiling
Forrester Research — The Automation Plateau
Typical Max Coverage
25%
Beyond this point, the manual effort required to maintain scripts consumes 100% of the team's capacity. The remaining 75% of the application surface defaults back to slow, manual verification.
Operational Waste Metrics
Per team member lost to fragmented toolchains and manual triage (GitLab).
QA budget redirecting to manual script maintenance (World Quality Report).
CI compute wasted on inconsistent tests (Google Engineering Data).
The CheckMate Solution
Time for me to say something good about my app. CheckMate transforms manual QA from a bottleneck into a high-speed feedback loop. Let me take away the busy work of setting up your testing and sending it out to your team. CheckMate centralises it all!
Reclaiming Testing Time
- Admin & Docs
- Test Execution
- Exploratory QA
By centralizing management, CheckMate converts 30% of administrative overhead into exploratory testing time, uncovering 25% more critical bugs.
Why CheckMate?
Rapid Creation
Reduce test design time by 40% with intuitive UI.
Centralized Repository
Eliminate "The Spreadsheet Mess" forever.
Automated Reporting
Get live statistics and automatically generated reports.
A Little Bit About Me
My Journey
I'm Ewan, a Brisbane based software engineer. It's at this point you'd expect me to say that testing is my passion. Hmm, maybe not. I feel I've made it quite clear my hatred for Excel sheets, thus CheckMate was my first real experience of looking at a problem and saying "I can do better than this". I've been a software engineer for 3 and half years. Wokring in smaller companies with no dedicated QA team, I've seen the distaste for testing first hand, whether that is setting up test cases, sending them out to the team, or especially clicking on an Excel link in Teams to have it open in the app or browser instead of the Excel app 😭.
I had an blast desiging and creating CheckMate, even as it consumed my life for 6 months. I have learnt A LOT. For the designing phase, I learnt that my vision of the API and endpoints did not match the reality of my Figma designs, with many endpoints being redundant, unecessary, or incomplete, meanwhile important ones were missing. For the development phase I learnt how to properly implement automated testing on my API, as well as the difficulties of translating my Figma designs into Angular components, don't even get me started about scope creep. Finally learning to host my backend through cloudflare tunnels and aquiring developer accounts so that Apple would leave me alone every time I tried to open my app. I achieved everything I set out to do and am looking forward to the next minor issue that I take personal offence to.
