It's Been a While...
Last summer, I landed a pretty amazing job. It was a needed transition - the kind of job that I wanted since college, but one that was difficult to find in Oklahoma.
The only issue is that it's a lot of work. I'm not complaining, but when I have a family with three pre-teens in every kind of sport and after school activity, along with an all-consuming writing hobby, I learned pretty quickly that I couldn't fit it all in. PitchStats ended up on the backburner.
That said, I didn't abandon PitchStats. I've been lurking...planning...building... 👀
A Labor of Love
For what it delivers, PitchStats never took an unreasonable amount of time, BUT it was never quick. For the first few events, downloading tweets and corresponding likes took hours...and I would often have to babysit the process as it ran because it could fail for various reasons and need to be restarted.
Once I had the tweets, I would process them. I usually reserved a weekend - not because it took two solid days, but because it did take several hours, and I often needed time without interruptions.
- I build the event, download hashtags from official websites, and scan the pitches to find additional tags that I would add on my own.
- I check for spam. Just about every pitch event trends, and twitter spammers love to pile on. Each event, I scan for spam and flag it.
- The most arduous part of the process is labeling Agents, Editors, and Publishers. I have scans that helped me identify them, but overall, it's painstaking. I step through profiles until I go cross-eyed, trying to differentiate whether or not the person liking tweets is an industry professional or an over-zealous supporter.
After processing the event, I would write them up. This part is fun, but it was painstaking. Getting just the right view, marking it up appropriately, and then communicating the message in Twitter's shorthand took quite a bit of time. I tried to post one thread a day, usually putting out 4 or 5 threads for a small event.
End to end, a PitchStats event could easily take 10-20 hours of work over the course of 7-10 days. While I personally enjoyed the results and learned a lot from the analysis, it's proven difficult to manage with my currently life.
Rather than walk away, I decided to retool.
- I improved my scripts for downloading tweets and likes drastically over time. While they still take several hours to run, they run in the cloud and rarely have errors.
- I wrote better processes for identifying spam and popular hashtags.
- I trained an Open.ai model to look at Twitter profile information and to tell me if someone is an Agent, Editor, or Publisher. It's not perfect, but it's surprisingly close - and a HUGE game changer.
- I built dashboards of standard analytics I usually look at for each event.
I didn't build all of this overnight. Instead, it's been a gradual work for the last six months or so.
Recently, I took my new process and tested it across three of the recent pitch events: IWSGPitch, KidLitPit, and PBPitch.
When I processed these events, I was able to do one in about an hour!
I processed IWSGPitch on a Friday night and the other two during an afternoon on Saturday. KidLitPit was a new event for me - I had never processed it before - and I was still able to knock it out quickly.
Completing the Puzzle
This brings me to the last piece: Presentation.
My prior tweet format won't work going forward. I don't have time to crop and edit countless images or to write/rewrite twitter sized tweets. As with everything else, I need to standardize and streamline the process.
At first, I hoped my dashboards would be a good way to disseminate the data, but while they work alright for sharing with individuals, they won't work for mass consumption.
Which brings me back to this blog. I started it with the hope of expanding the PitchStats platform, featuring anecdotes or statistics that didn't fit the general twitter format. I didn't have much time or inspiration for that, so it sat dormant, but now, I'm hoping it will be the perfect platform for what I need.
I'm going to start experimenting. To test out my ideas, my first post will be focused on last fall's MoodPitch. It's pertinent because it was a very successful pitch event, and the next event is coming up. My hope is that the blog format will prove to be even better for authors. I hope to allow you to navigate the data directly and understand what agents and industry professionals are looking for.
I'm putting the final touches on the MoodPitch post. It should be live in the next day or two. I'll be looking for feedback, so please don't hesitate to let me know what you think.
- Does this format work for you?
- Is the right data presented? Is it interesting/useful?
- Does everything make sense?
- Is there too much information? Should it be broken up into several posts?
My hope is that this will be a better way to communicate my analysis to the writing community - but it could also end up being lost to the internet. If this is something you enjoy, please let me know.
As always, if you'd like to support my efforts, please like and retweet the #PitchStats posts...and consider subscribing below. 🙃