Issue #45: Holding customers' hands 🧑‍🤝‍🧑

An onboarding flow to help them see the value fast

First impressions matter. 😁 

Table of Contents

Onboarding

Up to a few days ago, if you signed up for HeyEcho, you would see a menu to create a workspace, generate a blog post, create the content hub, etc. More or less, it was like throwing someone in the middle of a pool, hoping for that person to learn how to swim. 😂 

That’s why I spent a few days creating an onboarding flow. The goal is to make the customers see the value with as little friction as possible.

It’s split into 3 main consecutive steps:

  1. workspace creation: a very simple input text area to insert the website,

  2. content hub generation: as soon as the workspace is created, the content hub is generated, and the user is asked to pick one topic to generate the first blog post,

  3. blog post generation: the first blog post is generated by showing the live progress, and once done, the user is asked to be redirected to the content library.

Anyway, a video is worth a thousand words, so check it out here!

HeyEcho’s blog post driving organic growth

This is so cool! 🚀 

NextCommit has a few other blog posts not generated by HeyEcho, and it’s exciting to see that these two are already contributing to driving the organic growth. Starting from Monday, I’ll resume using HeyEcho more consistently for both HeyEcho and NextCommit.

Blog performance first improvements

Last week I shared how bad the performance is for HeyEcho’s blog, and it’s still a work in progress, but it looks way better already! 🔥 

Lighthouse score for performance improved:

  • on mobile: from 65 to 91,

  • on desktop: from 77 to 98.

Not bad, huh? 😎 

What did I address? A few things.

  1. building for production: building the website with NODE_ENV=production. Yeah… what a shame that I was not doing it already… 🤦 

  2. enable text compression: I enabled gzip and brotli on the ingress load balancer.

What else am I trying to address?

  1. http2: I’m trying to enable http2 as suggested by Lighthouse, but I’m having some trouble enabling it. I’m still missing something in how to do it in my setup that relies on k8s Nginx Ingress Controller backed by a Load Balancer on DigitalOcean provisioned using Pulumi. 😅 

  2. reducing the JS/CSS bundles: this is tricky because I’m not an expert and my setup is not straightforward. I have a monorepo with EVERYTHING: infra with Pulumi, API with the python server, and the frontend with the website and the webapp. The frontend itself is implemented using turborepo in a way that I have a bunch of shared packages that are dependencies of the two applications (website and webapp). I had a few issues when trying to tune the building process, like bundled code that was missing dependencies at runtime, etc., and yeah, both Claude3.7 and GPT4o struggled as well. 😅 

  3. properly size images: currently, I upload to the CDN the hero image of each blog post, but for example, for mobile, this is a bit on the big side, hence affecting the performance.

  4. blog posts index pagination: this is related to the previous point, that is, currently the index of the blog is not paginated, so with 20+ posts published, this means also 20+ images to download.

Conclusion

Feature freeze! 🥶 

I think that the current version of HeyEcho has everything needed to be valuable to anyone that wants to start investing or increase the organic growth through a blog-first content marketing strategy by saving a lot of time without compromising on quality.

In the next weeks, I’ll start again reaching out to people and launching HeyEcho on different directories. Time to wear the sales guy hat. 😁 

If you’re interested in following my journey, make sure to subscribe or follow me on Bluesky, X, and LinkedIn

Reply

or to participate.