Competitors Monitoring
Competitors Monitoring
TL;DR
Competitor research at Elastic Email happened often, but rarely in a coordinated or repeatable way. Different teams checked competitors when a project demanded it, then the knowledge quickly became scattered or outdated.
In 2026 I set up a recurring monitoring process covering around 20 competitors and roughly 60 pages. I first tried to manage it manually, then with page monitoring tools, before eventually automating most of the process with Claude. The result was a lightweight research system that could produce regular competitor updates with relatively little manual work.
My role: full ownership
The problem
We were keeping an eye on competitors, but not in a systematic way.
Knowledge was scattered
Different people checked competitors at different times and for different reasons. Useful findings often stayed with the person or project that produced them.
Changes were easy to miss
We wanted to know when competitors changed pricing, messaging, features or their broader business direction. Without regular monitoring, important changes could go unnoticed for many weeks or even months.
Manual research did not scale
Email marketing and email API are crowded markets. Once I listed the companies worth following, I had around 20 competitors and 60 relevant pages to monitor. Checking them all manually was less than realistic.
What I did
Defining what to monitor
I started by building a list of competitors and deciding which pages were most useful to track.
For most companies this included:
I was particularly interested in pricing changes, messaging shifts, new features and larger business developments.
Starting manually
My first approach was simply to review the pages one by one and record notable changes.
It quickly became obvious that this would take too much time. With dozens of pages to inspect, the monitoring process itself risked becoming a full-time task.
Testing page monitoring tools
I tried Distill to automate comparisons between different versions of a page.
This helped identify where changes had happened, but I still needed to inspect every result manually, decide whether the change mattered and write it down.
The process was more efficient, but still too labour-intensive.
Automating the research with Claude
The biggest improvement came when I started using Claude Cowork.
I created scheduled tasks that ran every week. Claude used Chrome to visit each competitor's website, review the selected pages and prepare a structured report describing relevant changes.
This shifted my role from collecting information to reviewing and synthesising it.
Outcome
The project did not lead to one major strategic decision or measurable business result.
Its value was operational.
I turned an inconsistent manual activity into a repeatable research process that could run largely on its own.
The final workflow combined automation with human review. AI handled the repetitive work of visiting pages and documenting changes, while I remained responsible for interpreting the findings, deciding what mattered and compiling the final reports.
For me, the most important outcome was proving that a research process with a relatively large monitoring scope could be maintained without requiring a proportional amount of manual effort.