6 Steps to build your Predictive Model

Options
José Forte
José Forte Posts: 9 ✭✭✭

If it's the first time you've heard about me or predictive models, risking ~650€ were already worth it.

Let me tell you the tale of a simple RevOps champ with a simple desire, to find a better way to make pipe predictions.

The Quest for the Perfect Prediction

At the begining of Q2 I had a mission - prove that it was possible to make reliable predictions of accounts inside a pipeline.

Let's face it, no one on the B2B, SaaS or Statup community is posting about it (probably because unlike me they knew that it's not an hot topic xD).

So, I did what any self-respecting data enthusiast would do: I went digging. Research reports, industry blogs, the whole nerd routine. But the insights just weren't there. Then, one bleary-eyed night after battling spreadsheets for what felt like an eternity, I stumbled upon something… unexpected.

The Unexpected Ally (Betting Forums)

Hold on, hold on, before you judge! These forums, filled with passionate fans dissecting every stat and player quirk, were a goldmine of real-world data disguised in funny memes and trash talk. It was like uncovering a secret garden of insights hidden within the jungle of the internet.

Here's the key highlights I got from it:

  1. Bigger complexity doesn't always mean better results
  2. Always test all variables
  3. Even the weirdest programming code can live on an spreadsheet if you understand the end goal
  4. You should not try to make money out of those predictions

Time to risk it all

After being able to pick several predictive model ideas and making my own model, something was missing. I needed real-world validation, to discuss it with similar piers and keep getting inspired.

So, I took a risk, testing my own model against the odds of Solverde.pt , ChatGPT and Google Gemini on a thriling stage: Euro Cup and Copa America in order to attract you to this topic.

1721008985871?e=1726704000&v=beta&t=UaPglb7Zi9rCDbU8Tb96uAreqHwVdcdjGRvm1Aw3Qj0

Euro2024 Prediction Analysis

1721013897185?e=1726704000&v=beta&t=5viIs5yo6tsFzfBQknZYy3luDJAws70eE77tSD9EqPc

Copa2024 Prediction Analysis

Victory and focus on key takeways

The results were astonishing! My model outperformed our AI friends, predicting a staggering 46% of Euro Cup outcomes (23/50 game winners/draws predicted and 11 final results forecasted) and a phenomenal 68% of Copa America results (21/30 game winners/draws predicted and only 2 end results forecasted).

But the real win wasn't the bragging rights (ok, part of it was!). But it was also taking this data magic and understanding how between the account data I got from 6sense and the powers of HubSpot could I make something but for work-life. With a few clicks and some formulas love, I automated customer value prediction. No more tedious manual analysis, just bam! Instant insights at my fingertips.

Sometimes, the best solutions come from unexpected places. By embracing calculated risks and unconventional data sources, we, as RevOps professionals, can revolutionize the way we understand and engage with our customers.

6 STEPS TO MAKE YOURS

Check the full article to understand how

IMPLEMENT IT IN YOUR CRM

Check the full article to understand how

FINAL REMARKS

This, my friends, was the story of a RevOps renegade who found inspiration in the most unexpected place. Proof that the key to customer insights might be lurking in the weirdest corners of the internet, disguised as bad sports puns and questionable life choices. So, the next time you're drowning in data, remember: the key to customer insights might be a click away from a passionate – and possibly slightly unhinged – online forum. Just don't tell your boss you learned it all from betting forums. They might not understand the power of a sassy data scientist with a gambling streak.

If somehow you had a good moment reading this article please show it with a lovely kudos, share with your network so we keep talking more about this topic and don't be shy to ask questions or to simply show off your success with the others!

Comments

  • @José Forte Such a great post! Thanks for sharing

  • Ami Arad
    Ami Arad Posts: 76 6senser

    Love it! If/when it comes time to rollout 6sense to your sellers, do NOT underestimate the value of gambling analogies. In my experience, both as a seller, and having trained 1000+ sellers on 6sense now, a lot of sellers consider themselves gamblers. Half or more of their compensation is a bet on themselves. As such, I frequently invoke gambling analogies when explaining the Model metrics report. 6sense predictions are not guaranteed; they just improve the sellers' odds they'll open an opportunity, and it's important they understand their odds if they use 6sense vs. ignoring it. I created a video about it here:

  • José Forte
    José Forte Posts: 9 ✭✭✭

    Amazing @Ami Arad! Don't think we've the Model report capability but really liked the win probability <> gambling methodology, will make something of my own and try it :)

  • Love the passion and relevant + relatable example. Thanks for sharing!