I'd only just finished building a system, using Make, that automatically piles readers up into a Brevo list, when a new thought hit me: "now that the list exists, it'd be nice to glance at it without having to log into Brevo every time." So I set it up to copy each new reader into a Google Sheet as well. It had more sticking points than last time and took about two to three hours: access permissions, avoiding duplicates, and a scenario that switched itself off in live use, and at one point I simplified the whole design. Here's another record of working through it screen by screen with an AI agent (Claude Code).
A spreadsheet and a reader list, linked by a two-way sync.
What I wanted to do
Brevo was dutifully building up the reader list, but logging in just to take a quick peek felt like more friction than it should be. If a new row appeared automatically in a familiar Google Sheet every time someone signed up, I could just open it and see the whole list at a glance.
Luckily, the Make scenario from last time (form submission → Webhook → add to Brevo) already existed, so in theory I just needed to bolt on one more block: "also write to a Google Sheet." In practice, I ran into a few walls I hadn't expected.
Snag #1: the Google permissions were way too broad
When I went to connect Make to a Google account, the confirmation screen asked whether to allow access to "all files in Google Drive" or "all spreadsheets."
I'd assumed I could scope this down to just the one spreadsheet I actually needed, but it turns out Make's Google Sheets integration doesn't offer that option — granting access would let it reach into other, unrelated files in my everyday personal Google account too.
That gave me pause, so I went ahead and created a brand-new Google account dedicated purely to automation, and moved the spreadsheet management over there. That way, the blast radius is limited to files inside that dedicated account. A bit more setup, but worth the peace of mind.
Snag #2: "avoiding duplicates" was harder than expected
Last time, I'd set Brevo to update an existing contact instead of throwing a duplicate error, so I tried to build the same "update if it exists, add if it doesn't" logic for the spreadsheet too.
The idea looked like this:
Webhook receives the data
└─→ Search the sheet by email address
├─→ Not found (new) → Add a new row
└─→ Found (existing) → Update that row
But the search module I'd picked turned out not to return "which row number this match was on," and a couple of times I ended up chaining two modules in sequence when I'd meant to branch them apart instead. One small configuration mismatch after another, and it took a lot longer than I expected.
Deciding to simplify, deliberately
I could have kept patching each issue one at a time, but at some point I stopped and asked: "do I actually need this much complexity?"
Once I framed it that way:
- The list that actually matters for sending emails lives in Brevo (which already handles duplicates correctly)
- The Google Sheet is really just a list for a human to glance at
—it became clear that the occasional duplicate row in the sheet wouldn't really cause any harm. So I dropped the whole search/branch/update setup and simplified it down to just always appending a new row. Cutting it down to a straight line by deleting two modules was a real relief.
Snag #3: a production error that auto-disabled the whole scenario
With the simpler version in place, testing went smoothly — but then a real submission from the live site triggered an "Internal Server Error," and Make automatically turned the whole scenario off in response.
My first reaction was panic, thinking data had vanished. But Make's "History" tab had a clear record of exactly when the error happened and which module it came from, along with the fact that the scenario had been deactivated as a result. It looked like a transient hiccup on Google's end — turning the scenario back on resolved it. Learning that Make has this auto-shutoff safety net was a useful discovery in itself.
Snag #4: the registration date kept showing up as the literal word "now"
Finally, I wanted each new row to record a timestamp automatically. I entered a date function, but instead of a computed date, the sheet just showed the literal text "now".
Even though I thought I was selecting a function from the little icons in the editor, it kept getting typed in as plain text, or the function ended up assembled slightly wrong. After a few rounds of trial and error, clearing the field completely and typing {{now}} directly from the keyboard finally did the trick — producing a real timestamp like 2026-06-22T10:42:47.765Z.
How it felt
This round had noticeably more obstacles than last time, and took about two to three hours. "Search, then branch" is a pretty basic idea in programming terms, but inside a no-code tool, figuring out how to actually wire the pieces together turned out to be the harder part — a fresh realization for me.
I said "this is too hard, I don't get it" more than once along the way, but sharing screenshots each time and getting help pinpointing exactly what was wrong — including the decision to back off and simplify rather than push through — made all the difference. Knowing when to step back from an overly complex approach felt like its own small win this time.
Checklist: what I did, step by step
- ✅ Noticed the overly broad Google permission request and created a dedicated automation-only Google account
- ✅ Attempted a "search → branch → update" setup, got stuck, and abandoned it
- ✅ Simplified to "always just append a new row," accepting occasional duplicate rows
- ✅ Discovered a production error auto-disabling the scenario, traced it via History, and re-enabled it
- ✅ Fixed the registration date showing up as the literal text "now" so it now records a real timestamp
- ⬜ Reformat the timestamp column into a more human-readable format
No-code automation tools are convenient, but even a seemingly simple requirement like "how do I handle duplicates" can take more effort than expected. When an overly clever setup gets stuck, it's worth pausing to ask whether it's really necessary in the first place.
Other books you might like
You're welcome to take a look.
EAST TO WEST: The Heart of Haiku, the Light of Scripture
The wisdom of pausing rather than forcing something overly clever — the kind of quiet reflection this collection finds through haiku. This is a book my father-in-law left behind.
View on AmazonCataract Surgery-A Memoir: Double cataract surgery in Osaka, Japan
Another new experience at 66 — an honest account of the cataract surgery I went through.
View on AmazonSushi Business in Britain: the complete four-volume series
A novel series following thirty-five years in the British sushi trade, from apprentice to factory manager to a small sushi school. The complete edition offers all four volumes at a better price.