If your business is producing any amount of data more than 200 rows per week (and all businesses are), then you need to perform deep-dives from time to time in order to understand your company’s performance and improve. Otherwise, you’re likely missing data points that would help you achieve step-functions in growth.
The human attention span and amount of data you can process with your eyes and brain are very limited. You should therefore focus on high level, condensed data. High-level data can tell you the full story of what’s going on and how you can improve in an easily understandable way.
Let me take a step back for a moment to analyze the term “deep dive”. I love spearfishing. Even though I’m a lousy diver, I try to do it every summer off of the beautiful shores of Turkey. I think “deep-dives” is a perfect analogy in business because:
A photo I took in one of my favorite spearfishing spots — Datca / Mugla / TR
- Just like your air when you are spear-fishing, your time is limited. It’s your ultimate finite resource
- You need to have a basic idea of what you’re looking for: a fish!
- You have to come back to the high level eventually; lungs will only last the average person around 3 minutes
- Water pressure makes life harder as you go deep, and data complexity acts in the same way. The answer? Better data normalization, formatting, grouping and visualizing.
Example 1 — Heatmaps and better use of splash screen for web pages
1 Sentence Learning: After collecting behavioral data for Crossover’s most critical web pages, we discovered we were not effectively using the page height based on page-scroll data on hotjar.
Action: We re-prioritized our content and moved mission-critical content to the top part of the pages. We also created two A/B experiments on the landing pages; right now we are trying content heavy and over-simplified versions simultaneously.
Example 2 — Testimonials & Conversion
1 Sentence Learning: We’ve been producing testimonial content based on hypotheses, but by leveraging Google Analytics data, we were able to understand the behavioral difference of our prospects before and after engaging with our testimonial content.
Action: We built an automated dashboard to track behavioral differences and the performance of our testimonial content.
Example 3 — Channel Comparison
2 Sentence Learning: Crossover is receiving over 17K applicants per week, so a deep-dive was necessary in order to compare paid and organic channel performance with a functional domain break down. This deep-dive revealed that organic channels are performing as well as paid channels.
Action: Action: Conducted a deep-dive into organic channels to discover the dollar value received through marketing content, social media, and organic search traffic.
Example 4 — Crossover Added Value in the Emerging Markets
3 Sentence Learning: As the inventor of the terms “Cloud Teams” and “Cloud Wages,” Crossover is offering massive value to professionals from emerging markets. We wanted to know exactly how much. We worked on four different models to understand Crossover’s added value in each market.
Action: Now that we know the arbitrage opportunity of our investments, we are using this data to prioritize markets and marketing investment.
Example 5 — Testing Content Pass Rates
3 Sentence Learning: At Crossover, we are evaluating 17K candidates on a weekly basis and our candidate testing process has 4 major parts: Basic Fit Questions, Cognitive Aptitude Test, Subject Matter Questions and Free Response Questions. After doing a deep-dive into candidate data, we discovered some of our pipelines were eliminating good candidates and some were letting bad ones pass. Considering the huge amount of money we’re investing in lead generation every year, these mistakes have dollar-value ramifications.
Action: Built a customizable report for internal Crossover teams, giving access to all candidate data, showing alarming pass rates on each step of testing with the ability to perform pipeline specific deep-dives when needed. As a result, we fixed the broken pipelines!
Remember, deep-dives are one of the highest impact activities you can do for your business. Only by doing deep-dives can you;
- Inspire others to do deep-dives with the data you have
- Apply improvements based on your learnings, which leads to process improvements for days, weeks, months and even years in some cases
- Have a better understanding of your own data, leading to you start asking better questions
One Important Note
I thought my first serious deep-dives were great, but my colleagues did not understand a thing, simply because I failed to tell a story with the data. I suggest you focus on what matters. What do you think would really improve the process and has dollar value?
You have to make sure your deep-dives are producing signals, not noise.
Signals versus Noise
Deep-dives can easily turn into data overkill. Whether you’re performing deep-dives for your boss, teammates, or subordinates, you need to guide yourself with some rules. Don’t let deep-dives turn into a distraction.
- Do not communicate deep-dive data worth less than 1% of your annual turnover to your boss
- Do not communicate deep-dive data worth less than 0.5% of your annual turnover to your peers
- Do not communicate deep-dive data worth less than 0.25% of your annual turnover to your subordinates
If you think we share the same mindset and are interested in working with us, feel free to browse our open positions.
Special thanks to Heather Aholt for editorial support.
Remote Jobs, Tools, and Thought Pieces
Make sure you are subscribed to WorkRemote, the only blog focused on remote work at scale. Some of the great content from WR:
- Daily Habits of a Successful Remote Manager
- What is a Class A remote job?
- How To Make a ‘Gold Standard’ Remote Work Office
- 2 Critical Steps for Fortune 500 Managers Going Remote
- How to Perform Shrink-to-Grow in a Remote Team
- How to Hire Your First Remote Employee
Thanks to Heather Aholt.
Want to read more?
We have a lot more where that came from
