Mage is a growing AI start-up company that went through many transitions during it’s time to establish an identity, direction and brand. I was hired as the first technical writer and expanded their content outreach and strategy for technical audiences primarily focusing on those that code. From writing posts to managing writers I was deeply involved with project management, content strategy, and developing a community. 3 month after working, I was promoted to hire and assemble the full-time writing team. This team would later become the DevRel/Documentation team where I shifted focus to growing our community and developing Mage Academy, a platform to learn about Mage and AI.
Highlighting the differences between the boards, SEO automations and constantly generated, so it would mostly be like finishing a checklist with some grouping. Mage Academy was owned by me, so I was able to plan out first and carefully weave in when would be a good time to schedule. The blog was a cluster of all content and resulted in the greedy coin problem of mass producing content, waiting for quality content, or perfecting content.
Project Management
The project management model I adopted followed Agile Scrum, focusing on giving writers independence and the ability to choose what they want to write about to be the most creative. I owned 3 AirTable boards: SEO Content Calendar, Mage Academy, and Blog Schedule. To speed up our processes, I also developed automations and learned new frameworks to best approach each board.
SEO (Search Engine Optimization)
Using ahrefs, lighthouse, and other SEO tools I was able to identity ways to increase our search ranking from a technical perspective. This would require coding changes to the blogs, metadata tagging, and optimization on various devices. This felt as a blast from the past as I went back to working on more frontend coding tasks and incrementally improving our SEO on the technical end and skyrocketing our Google search results. I’m especially proud of my work on this. By the end of the optimization, a search for Mage would show our company instead of the WoW class on the front page. This heavily expanded our reach as our blogs were able to compete and outshine Medium domains and various competitors.
Mage Academy
Mage Academy was my first step into Instructional Course Design. It focuses on multiple personas for Data Scientists, Data Analysts, Data Engineers, and Software Developers. Mage Academy had a huge emphasis on data science, machine learning, and using the product for those specific end users. It spanned dthrough the Machine Learning Lifecycle and it’s purpose was to show how Mage is a no or low-code tool that can help with the roles and responsibilities of an end-to-end ML team. There were modules written for Data Processing, Data Cleaning, Model Training, and Model Deployment, each resonating with a specific job role.
Blog Schedule
The entire team of growth/marketing owned this board. This was by far the most volatile board, which led to a nightmare in managing processes. Pushing deadlines, cancelations, and scheduling changes made us seriously reevaluate our bandwidth as the writing team was stretched thin across managing workers in other timezones. It made me appreciate Sprint Retrospectives and over time we were able to find our rhythm and ideal posting schedules. Admittedly, this left burnout through the ranks, and our grwoth team’s biggest battlefield. There was a set cadence to beat, a timeline due to trends, and and priority would flip as fast as they were determined as we built in public.
Content Strategy
As part of our content strategy, I tried out many different kinds of content, owned analytics dashboards using Amplitude, and after a lot of trial and error found out what worked, what didn’t, and directly reported the results to the CEO. Starting off, we went for inorganic methods to grow our brand, these would include guest posting on higher ranked pages, joining communities and posting there like IndieHacker, Dev.to, and other tech or learning focused platforms. After growing our top funnel, I went back to focus on retention methods for our existing customeres, such as well written documentation and support sales staff.
Posts
Posts were intended for technical and non-technical audiences, but had a marketing and educational aspect of it.
- Posted Mage focused articles in my Medium in the Mage Publication.
- Guest Post on Simple Programmer
- Guest Post on The AI Journal.
- Cross posted all technical articles in Dev.to and IH as well.
Documentation
I wrote documentation for functions, sample dataset labels, and third-party integrations. You’ll find these docs at https://docs.mage.ai/. These docs were used to create blog articles, video tutorials, and API documentation
I owned scripts and recording, but for some I also presented videos on datasource integrations with Mage. Available on YouTube. See most video before “Mager Transformation”.
There were more videos in the past, but Mage no longer supports those features (Snowflake, and others datasources), so they were taken down.
Short-Form
Short-form content took a lot of effort to get right and would nuke our metrics often. When Mage was small and didn’t have a well established brand and tried many different types of short-form content to quickly prototype how our audience would react to it. In the end, we found Twitter as our top developer platform and began to shift content efforts to drive Twitter engagement. During this time, Technical Writers were given roles similar to DevRel and everyone in the company was to like, comment, share, etc… when a post was made.
On Twitter we focused on messages for
- iPhone13 Giveaway/Raffle
- Release notes and launch videos
- Tweeting out investors and established individuals
Community
Speaker at Technical Events
I presented at TreeHacks 2020 to do a livestream of the Mage API/Website demo and answer questions. Mage was a Hackathon challenge sponsor and had their own category with cash prizes. During the presentation, I walked through end-to-end how to transform data and (re)train a machine learning model with Mage to optimize it for a given use case. I wrapped up by answering questions from the audience and explaining our Hackathon challenge and prizes.
Discord Community Server
I pitched, developed, and moderated our Discord Community Server for Mage. Inside, there were founders, students, and staff ready to answer product questions and teach more about data science. To keep the server on brand, I worked with our designers to develop “level up” roles, hidden channels, and bots to track metrics. I setup webhooks to sync our Twitter, AirTable, and Slack to also inform the Discord server in real-time as we sent out more releases.