profile

Shachar Meir

Why Self-service Analytics Keeps Failing (AI doesn't fix it)

Published about 11 hours ago • 1 min read

Hi Reader,

AI makes self-service analytics look inevitable.

The pitch sounds something like this:

→ Plug an AI agent into a clean semantic layer
→ Add good metadata and context
→ Let stakeholders ask questions in plain English
→ Done

Everyone gets answers instantly.
No dependency on data teams.

Sounds great.

But here’s the problem:

Self-service analytics has been “almost working” for over 20 years.​
And success rates are still… disappointing.

Why?

Because it quietly relies on a set of assumptions:

  1. Stakeholders know what they actually want
  2. Stakeholders can express it clearly in words
  3. Words always mean the same thing to everyone
  4. Stakeholders understand the data well enough and want to be self-served
  5. Stakeholders even need that level of flexibility in how they consume data

If you’ve worked in data, you know how shaky these assumptions are.

AI makes the interface better.
It doesn’t fix the underlying problem.

I’m not saying it won’t work at all.

But expectations need to be… realistic.

In this mentoring conversation with Omar Brid, we break this down properly:

→ where self-service actually works
→ where it fails (and why)
→ what data teams should do instead

Watch it here 👇

Why Self-Service Analytics Keeps Failing Why Self-Service Analytics Keeps Failing [30:19]
Why self-service keeps failing, and how to make it work?

Have a great weekend!

Regards,
Shachar

P.S. I've recently launched my website – if you’re serious about growing beyond execution work and into real impact, it’s worth a look.

♻️ Please share these links with 2 colleagues or friends to help them grow:

Shachar Meir

Data Advisor

Read more from Shachar Meir

Hi Reader, Most people think "personal brand" means LinkedIn, YouTube, TikTok.It doesn’t. In fact, every one of us has a personal brand. Inside your company for example, there are people who: trust you on certain topics come to you for advice see you as the “go-to” for some things That’s your brand. The real question is:→ are you shaping it intentionally, or leaving it to chance? When Miquel asked me how to build a personal brand, I broke it down into 4 simple questions: What’s your goal? Who...

7 days ago • 1 min read

Hi Reader, People think doing more = faster growth.It doesn't.In fact – in many cases it's the opposite.Recently I recorded a mentoring session with Sherin, a Data Analyst at Booking.com.She felt stuck at Level 2 and wasn’t sure how to grow. So I asked her what she works on.She said: “Everything. Finance, sales, marketing…” That’s the problem. When you spread yourself across too many domains: you never go deep enough you never build real context and expertise you never become the go-person...

14 days ago • 1 min read

Hi Reader, Most data professionals think they’re doing a good job. They deliver what’s asked.They answer questions.They build dashboards. And that’s exactly the problem.They’ve become order takers. This used to be “fine”.Today? It's very risky. Because order takers: Don’t get promoted Don’t get recognition Are the easiest to replace (by AI or by someone cheaper) The people who grow fast do something different.They lead. Not with authority.Not with title.But with how they think and how they...

20 days ago • 1 min read