Simon Jackson retweeted
With apologies to Jay-Z, we'll just say that "a BI tool ain't one." We have soooo many analytical tools in our toolbox. Which one is "best" for any given ask? @drsimonj from Hypergrowth Data joined us to try to figure out an answer. analyticshour.io/2024/05/28/…
If you're not making decisions with data, don't expect to stay in the game.
The Moneyball story nails it:
Data creates decision leverage.
It can maximise the value of your resources.
If a competitor figures out how to do it,
You better get onboard or be knocked outta the game.
246
Struggling to come up with product growth ideas?
There are four dead-simple steps anyone can run.
All you need is your data and some thinking time.
Yesterday I shared how to generate lots of great test ideas, fast!
Missed it?
Get it here ↓
buff.ly/4as7L9s
279
Anyone can quickly design amazing test ideas based on their data
How?
1. Find undesirable patterns/outliers
2. Come up with explanations
3. Cut out any easily debunked
4. Design tests
Tomorrow I'll show my new site visitors how to do these
Subscribe at buff.ly/43AUNUL
219
Stop expecting data teams to make "recommendations."
Why?
"Recommendation" implies better or worse options.
But most decision options are about trade-offs.
If you can't commit to specific trade-offs up front,
Better to ask for "comparisons" or "evaluations."
304
Data has no inherent value.
It must be...
- collected with purpose
- stored with care
- managed effectively
- connected to the right problems
- The. List. Goes. On!
What customers DO with your products will always be more important than what they say.
Watch and learn.
243
What return should I expect on my investment in experimentation?
Time.
Think about it.
Running experiments doesn't get you customers or revenue.
They help you learn what's truly valuable, faster than almost any other approach.
2
1
592
You don't NEED elaborate hypotheses every experiment.
If testing is cheap, go test!
What IS important:
🎯 Call shots
Pre-declare your decisions given results.
🔍Be efficient
Define success consistently and invest in hypotheses to explore larger problem/solution spaces.
1
278
Too many data teams fail to deliver value.
Why? My observation...
Too few non-data people know how to share the problem that actually needs solving.
Not enough data people know how to elicit those problems.
It's like buying a bunch of hammers and hoping someone finds nails.
1
4
426
People don't want dashboards.
They don't want statistics.
They don't want reports.
They want business outcomes.
Data ain't dollars!
Very few data scientists know how to turn a failed promotion into an advantage.
Of the 10s of people I've seen fail, here's how the best respond and succeed:
1
3
1,024
Summary of how to turn a failed promotion into an advantage:
1. Pause
2. Own it
3. Find out why
4. Focus your energy
5. Get proof
1
1
355

