Building Grammarly Insights: Part II
Six lessons learned from managing this program, plus why you can't underestimate positive sentiment as a North Star metric.
This is the second post of a series looking back at the development and lessons learned from launching and managing Grammarly Insights. (Read part I)
Now that we covered the backstory and reason for building Grammarly Insights in the first place, I’d like to share a lot of the secrets to success in the hopes that some of these principles, frameworks, and practices might unlock actionable insights for you and your teams.
Lesson #1:
Ditch Development Cycle Conventions
Depending on the ambition behind your concept, your MVP may need a lot of time to bake in the oven. More so than what might fit into your typical comfort zone.
That was the case for us as we spent 4 months in design, with the revision count likely in the 20s-30s. And then we spent another 2 months doing coding and QA, given the complexity of the backend and data pipeline.
This isn't just a case of us being perfectionists. There were times we took breaks (see Lesson #2). We just really wanted to make sure we were at a point where the concept truly reinforced our core product & branding goals and nothing was confusing or unintentionally disempowering.
Otherwise, we could go to market and get a false negative reaction and lose the wind in our sails. In other words, sometimes people move forward too fast with an MVP that isn’t a good MVP at all.
Here’s the birds-eye view of the development timeline from our 1.0 launch:
Lesson #2:
Embrace Radical Collaboration
By radical calibration, I don’t mean inviting a lot of cooks to the party. Here’s a little story that happened while building our 1.0…
Getting our MVP and core concepts designed & thought out was mostly done in tandem (and in a silo) between two people: me and our lead email designer, Yura Tanskiy. However, after many rounds of design, we were both seeing things with tired eyes and didn't know how we felt about the concept anymore. We needed a break and some fresh thinking.
So, we did something pretty atypical in my experience. We gave up our baby to another tandem: a senior designer and PM from the product team.
We let them do their thing in their own silo for a few weeks, and they came back to us with a bunch of great ideas on how to improve both the look and feel and some of the main sections of content.
More importantly, they didn’t throw out the core of our thinking, which could have been emotionally devastating to our confidence. They simply got the lid off the jar that we had already loosened up.
It can be a growth edge for people to let go and kill your darlings, and it obviously could backfire in certain cultures or with certain personality types. But my suggestion is to try it sometime and you may be pleased with the results. And to not let your ego be your guide.
TL;DR: When you are an early-stage startup and trying to create something that has the potential to drive a huge impact, it’s not a bad idea to phone some friends and take a little time out.
For us, this was such a necessary and fruitful step for our journey in getting an MVP off the ground that was polished and made us proud.
Lesson #3:
Create Feedback Loops
In tech, the most passionate and urgent feedback eventually makes its way to the person most influential on a channel or project. But, I recommend going above and beyond the norm and proactively setting up a simple system for monitoring your program and encouraging over-communication internally.
Feedback for Grammarly Insights is multi-channel. We get more positive and constructive feedback on social channels than we do via users who interact directly with our direct ticketing system. Sometimes people just want to ‘vent’ or share out in the wild and are not necessarily expecting a brand to listen and take note.
But the compounding interest on being proactive and taking note can be a beautiful thing. Additionally, it can be very uplifting to see mostly positive feedback on a weekly basis for something you worked very hard on.
Here’s a simple flow-chart to illustrate those incremental monitoring systems. This is all made possibly by having a dedicated slack channel that is a shared responsibility by the support and marketing teams.
Lesson #4:
Iterate in Macro & Micro Cycles (like a product)
I’ll keep this one short and sweet.
Treat👏Email👏Like👏a👏Product.
In other words, you should iterate as quickly as possible based on data and feedback-driven insights. Create backlogs. Prioritize those backlogs using smart frameworks. Run experiments. Do all those things.
But eventually, you need to adjust your foundation to something new and improved that introduces additional, scalable ideas and features. Sometimes you may launch backend-only improvements, but hopefully, you can continue to iterate and breathe some lift in micro and macro doses so that your shelf life keeps improving.
For example, Grammarly Insights went from a 1.0 to a 2.0 in ~ 2 years.
Lesson #5:
Don’t Underestimate Qualitative Results
Grammarly Insights has been a workhorse for us in the sentiment category. It’s not uncommon to see testimony like the one below on a regular basis across social channels and especially on our Google Store Reviews for our Chrome Extension.
And this is validating and fun to share internally, but what should you do about it?
Well, why not quantify the qualitative? We did just that to get a better idea of the impact that Grammarly Insights drives and we compared it to our other email programs like Product Updates, Promo Offers, and our Blog Newsletter.
Here’s how Grammarly Insights stacks up on the quantitative and the qualitative:
Keep reading with a 7-day free trial
Subscribe to Scaling CRM to keep reading this post and get 7 days of free access to the full post archives.