"I Posted Image Tweets for 3 Days — Here's What the Data Says"

Disclosure: I work on Opennomos.

I'm 17 days into building OpenNomos Json — a tiny developer toolset with a Timestamp Converter and a JSON Parser. Last week we started an experiment: what happens when you add product screenshots to your build-in-public tweets?

Here's the 3-day data.

The setup

The experiment is simple: post one BIP tweet per day with a product screenshot, track views/likes/replies/reposts. Compare against our baseline text-only tweets.

The product: OpenNomos Json — two tools, ~10 daily active uses, early-stage SEO.

The screenshots: dashboard view showing both tools, logged-in state, real user context.

Day 1 (July 4)

• Tweet: First image tweet. Screenshot of the dashboard, "Day 15 building in public" narrative.
• Result: 2 views, 0 likes, 0 replies.
• Take: Terrible. Image didn't help at all. Could be content, timing, or algorithm.

Day 2 (July 5)

• Tweet: Second image tweet. Different angle — focused on the JSON Parser tool specifically.
• Result: Collected, pending 24h maturation.

Day 3 (July 6)

• Tweet: "Day 17 of OpenNomos Json" — screenshot of the full tool dashboard, sharing 3 specific learnings from 17 days of building.
• Result: Pending.

Early patterns

Image tweets aren't magic. They don't automatically get more engagement just because they have a picture. Day 1 was our worst-performing tweet in a week.

Quote tweets outperform everything. Our Quote Tweet experiment (quoting smaller accounts with relevant takes) consistently gets 3-6x more views than our main tweets. One Quote Tweet got 6 views vs 0 on the main tweet.

Thread decay is real. A 4-tweet thread sees views drop 10→4→3→2 across tweets. We're now testing "independent sub-tweets" — each tweet readable on its own, no "1/4" numbering.

What I actually learned about image tweets

1. Context matters more than the image. A screenshot of a dashboard people don't know means nothing. The image needs to communicate value without reading the text.

2. Dev tool screenshots are hard. Unlike SaaS dashboards (colorful, full of data), a two-tool developer utility looks... simple. The visual story isn't strong.

3. Small accounts benefit from interaction, not media. At 0-50 followers, adding images doesn't change reach. Engaging with other accounts does.

What's working better

The real growth levers we've found so far:

• Quote Tweeting smaller accounts: 3-6x views vs main tweets
• BIP (Build in Public) community: Consistent but small reach
• Cross-platform (dev.to): Longer shelf life, SEO value
• Consistency: Daily posting builds slow trust

The plan going forward

We're not giving up on images — we're giving them more time. Three days isn't enough data. We're also launching a cross-community repost experiment this week to test if the same tweet in two Twitter communities gets incremental or duplicate reach.

Bottom line

If you're building a developer tool and posting about it: start with text. Master the narrative first. Add images later, and only when the image itself communicates something the text can't.

Building OpenNomos Json in public. Follow along at json.opennomos.com.