Apr 12, 2022 · 23m · mad

Unlocking Data Observability with Monte Carlo's Barr Moses

Barr Moses · 18m spoken Matt Turck · 3m spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

In this fireside chat from Data Driven NYC, Barr Moses, Founder and CEO of Monte Carlo, joins host Matt Turck to explore the concept of data downtime and the discipline of data observability. Moses details Monte Carlo's technical architecture, the five pillars of data health, decentralized Data Mesh paradigms, and strategies for ensuring data reliability in modern enterprises.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 15.7% of the talking time here. How this is scored →

Matt as informed peer 2.7 Guest teaching 3.7 Guest disagreement 0.4 Matt pushing back 0.9
05100:0010:0020:000:09–2:28 · Matt as informed peer 1/10 Defining the Concept of Data Downtime Matt opens with a warm conversational question referencing Barr's Twitter handle to prompt the definition of data downtime. Barr provides a foundational overview based on extensive conversations with data teams.2:28–6:35 · Matt as informed peer 2/10 Real-World Consequences and Examples of Bad Data Matt prompts for real-world anecdotes where data errors had severe impacts. Barr details high-stakes examples involving public earnings, Fox live streaming, and Vimeo.6:35–10:02 · Matt as informed peer 3/10 The Five Pillars of Data Observability Explained Matt introduces a technical question from the community regarding data types and observability modes. Barr articulates the core five pillars framework clearly and thoroughly.10:02–12:47 · Matt as informed peer 3/10 Human Factors, Organizational Ownership, and Data Mesh Matt raises organizational dynamics, the data mesh movement, and buyer personas. Barr explains how organizational bottlenecks occur and how Monte Carlo fits into enterprise workflows.12:47–16:26 · Matt as informed peer 4/10 Product Tour: Architecture, ML Anomaly Detection, and Alerting Matt mildly challenges the reliability of machine learning anomaly detection, noting that ML is an imperfect science prone to noise. Barr agrees and clarifies how automation combines with contextual alerting.16:26–19:15 · Matt as informed peer 2/10 Data Discovery, Automated Cataloging, and System Insights Matt asks about product features like data cataloging and insights. Barr outlines key asset identification and Monte Carlo's philosophy on data discovery.19:15–22:31 · Matt as informed peer 4/10 Machine Automation versus Human Business Knowledge Matt brings up user questions on semantic knowledge and asks directly about Monte Carlo's defensible moat in a hot competitive market. Barr firmly rejects the idea that automated tools can replace business context before detailing their competitive edge.0:09–2:28 · Guest teaching 3/10 Defining the Concept of Data Downtime Matt opens with a warm conversational question referencing Barr's Twitter handle to prompt the definition of data downtime. Barr provides a foundational overview based on extensive conversations with data teams.2:28–6:35 · Guest teaching 4/10 Real-World Consequences and Examples of Bad Data Matt prompts for real-world anecdotes where data errors had severe impacts. Barr details high-stakes examples involving public earnings, Fox live streaming, and Vimeo.6:35–10:02 · Guest teaching 5/10 The Five Pillars of Data Observability Explained Matt introduces a technical question from the community regarding data types and observability modes. Barr articulates the core five pillars framework clearly and thoroughly.10:02–12:47 · Guest teaching 4/10 Human Factors, Organizational Ownership, and Data Mesh Matt raises organizational dynamics, the data mesh movement, and buyer personas. Barr explains how organizational bottlenecks occur and how Monte Carlo fits into enterprise workflows.12:47–16:26 · Guest teaching 3/10 Product Tour: Architecture, ML Anomaly Detection, and Alerting Matt mildly challenges the reliability of machine learning anomaly detection, noting that ML is an imperfect science prone to noise. Barr agrees and clarifies how automation combines with contextual alerting.16:26–19:15 · Guest teaching 4/10 Data Discovery, Automated Cataloging, and System Insights Matt asks about product features like data cataloging and insights. Barr outlines key asset identification and Monte Carlo's philosophy on data discovery.19:15–22:31 · Guest teaching 3/10 Machine Automation versus Human Business Knowledge Matt brings up user questions on semantic knowledge and asks directly about Monte Carlo's defensible moat in a hot competitive market. Barr firmly rejects the idea that automated tools can replace business context before detailing their competitive edge.0:09–2:28 · Guest disagreement 0/10 Defining the Concept of Data Downtime Matt opens with a warm conversational question referencing Barr's Twitter handle to prompt the definition of data downtime. Barr provides a foundational overview based on extensive conversations with data teams.2:28–6:35 · Guest disagreement 0/10 Real-World Consequences and Examples of Bad Data Matt prompts for real-world anecdotes where data errors had severe impacts. Barr details high-stakes examples involving public earnings, Fox live streaming, and Vimeo.6:35–10:02 · Guest disagreement 0/10 The Five Pillars of Data Observability Explained Matt introduces a technical question from the community regarding data types and observability modes. Barr articulates the core five pillars framework clearly and thoroughly.10:02–12:47 · Guest disagreement 0/10 Human Factors, Organizational Ownership, and Data Mesh Matt raises organizational dynamics, the data mesh movement, and buyer personas. Barr explains how organizational bottlenecks occur and how Monte Carlo fits into enterprise workflows.12:47–16:26 · Guest disagreement 1/10 Product Tour: Architecture, ML Anomaly Detection, and Alerting Matt mildly challenges the reliability of machine learning anomaly detection, noting that ML is an imperfect science prone to noise. Barr agrees and clarifies how automation combines with contextual alerting.16:26–19:15 · Guest disagreement 0/10 Data Discovery, Automated Cataloging, and System Insights Matt asks about product features like data cataloging and insights. Barr outlines key asset identification and Monte Carlo's philosophy on data discovery.19:15–22:31 · Guest disagreement 2/10 Machine Automation versus Human Business Knowledge Matt brings up user questions on semantic knowledge and asks directly about Monte Carlo's defensible moat in a hot competitive market. Barr firmly rejects the idea that automated tools can replace business context before detailing their competitive edge.0:09–2:28 · Matt pushing back 0/10 Defining the Concept of Data Downtime Matt opens with a warm conversational question referencing Barr's Twitter handle to prompt the definition of data downtime. Barr provides a foundational overview based on extensive conversations with data teams.2:28–6:35 · Matt pushing back 0/10 Real-World Consequences and Examples of Bad Data Matt prompts for real-world anecdotes where data errors had severe impacts. Barr details high-stakes examples involving public earnings, Fox live streaming, and Vimeo.6:35–10:02 · Matt pushing back 0/10 The Five Pillars of Data Observability Explained Matt introduces a technical question from the community regarding data types and observability modes. Barr articulates the core five pillars framework clearly and thoroughly.10:02–12:47 · Matt pushing back 0/10 Human Factors, Organizational Ownership, and Data Mesh Matt raises organizational dynamics, the data mesh movement, and buyer personas. Barr explains how organizational bottlenecks occur and how Monte Carlo fits into enterprise workflows.12:47–16:26 · Matt pushing back 3/10 Product Tour: Architecture, ML Anomaly Detection, and Alerting Matt mildly challenges the reliability of machine learning anomaly detection, noting that ML is an imperfect science prone to noise. Barr agrees and clarifies how automation combines with contextual alerting.16:26–19:15 · Matt pushing back 0/10 Data Discovery, Automated Cataloging, and System Insights Matt asks about product features like data cataloging and insights. Barr outlines key asset identification and Monte Carlo's philosophy on data discovery.19:15–22:31 · Matt pushing back 3/10 Machine Automation versus Human Business Knowledge Matt brings up user questions on semantic knowledge and asks directly about Monte Carlo's defensible moat in a hot competitive market. Barr firmly rejects the idea that automated tools can replace business context before detailing their competitive edge.

speaking balance: gold is Matt, purple is the guest (3 minute bins)

0:00 · Matt 20.2% · guest 79.8%0:00 · Matt 20.2% · guest 79.8%3:00 · Matt 3.9% · guest 96.1%3:00 · Matt 3.9% · guest 96.1%6:00 · Matt 10.7% · guest 89.3%6:00 · Matt 10.7% · guest 89.3%9:00 · Matt 20.9% · guest 79.1%9:00 · Matt 20.9% · guest 79.1%12:00 · Matt 12.4% · guest 87.6%12:00 · Matt 12.4% · guest 87.6%15:00 · Matt 21.4% · guest 78.6%15:00 · Matt 21.4% · guest 78.6%18:00 · Matt 15.4% · guest 84.6%18:00 · Matt 15.4% · guest 84.6%21:00 · Matt 23.7% · guest 76.3%21:00 · Matt 23.7% · guest 76.3%
Sharpest disagreement ▶ 19:26 Firm rejection of AI replacing business context

Barr immediately and firmly dismisses the premise of the question, stating unequivocally that machines cannot infer semantic business knowledge without human context.

Hardest push from Matt ▶ 15:13 Challenging machine learning imperfections

Matt pushes back on pure automated detection, pointing out that machine learning is an imperfect science that often creates false positives and alert fatigue.

Biggest teaching moment ▶ 6:36 Defining the core pillars of observability

Barr breaks down the conceptual framework derived from software engineering and systematically educates the host on how data observability applies to data engineering pipelines.

Matt holds his own ▶ 20:35 Drilling on market differentiation and moats

Matt displays deep market intelligence by framing the space as crowded and challenging Barr to explain Monte Carlo's durable advantage over copycat entrants.

the scores for every segment, with the reasoning behind each
ChapterTopicMatt as informed peerGuest teachingGuest disagreementMatt pushing backWhy
Defining the Concept of Data Downtime 1300 Matt opens with a warm conversational question referencing Barr's Twitter handle to prompt the definition of data downtime. Barr provides a foundational overview based on extensive conversations with data teams.
Real-World Consequences and Examples of Bad Data 2400 Matt prompts for real-world anecdotes where data errors had severe impacts. Barr details high-stakes examples involving public earnings, Fox live streaming, and Vimeo.
The Five Pillars of Data Observability Explained 3500 Matt introduces a technical question from the community regarding data types and observability modes. Barr articulates the core five pillars framework clearly and thoroughly.
Human Factors, Organizational Ownership, and Data Mesh 3400 Matt raises organizational dynamics, the data mesh movement, and buyer personas. Barr explains how organizational bottlenecks occur and how Monte Carlo fits into enterprise workflows.
Product Tour: Architecture, ML Anomaly Detection, and Alerting 4313 Matt mildly challenges the reliability of machine learning anomaly detection, noting that ML is an imperfect science prone to noise. Barr agrees and clarifies how automation combines with contextual alerting.
Data Discovery, Automated Cataloging, and System Insights 2400 Matt asks about product features like data cataloging and insights. Barr outlines key asset identification and Monte Carlo's philosophy on data discovery.
Machine Automation versus Human Business Knowledge 4323 Matt brings up user questions on semantic knowledge and asks directly about Monte Carlo's defensible moat in a hot competitive market. Barr firmly rejects the idea that automated tools can replace business context before detailing their competitive edge.

Statements from this episode (10)

Assertion Not checkable as stated
Moses: Public companies accidentally report incorrect financial numbers to Wall Street
“Companies actually report numbers to the street and accidentally report, report the wrong numbers, or about to report the wrong numbers.”
Barr Moses Apr 12, 2022 ▶ 2:42
Assertion Partly supported
Moses cites 2022 revenues for BigQuery, Snowflake, and Databricks
“So you're seeing companies like BigQuery with 1.5 billion dollars in revenue, Snowflake with a billion dollars in revenue Databricks with eight hundred million in accelerating.”
Barr Moses Apr 12, 2022 ▶ 4:42
Assertion Not checkable as stated
Moses: Schema changes are a major culprit for data downtime
“So actually schema changes are a big culprit for data downtime.”
Barr Moses Apr 12, 2022 ▶ 7:41
Insight
Moses: Validating data at a single pipeline point is no longer sufficient
“And so making sure that your data is accurate at only one point of the pipeline is just no longer sufficient”
Barr Moses Apr 12, 2022 ▶ 9:52
Insight
Moses: Centralized data teams become bottlenecks as data usage scales
“Today you have like hundreds of people working with the data. It does not make sense anymore that there's one team that sort of has the keys to it and is really actually just ends up as a bottleneck.”
Barr Moses Apr 12, 2022 ▶ 11:07
Assertion Contradicted
Moses: Monte Carlo uniquely offers out-of-the-box cross-system observability
“So we're actually the only sort of product and market that you can connect today to those sort of different systems and sort of automatically out of the box, get an overview of what the health of your data looks like and sort of observability for your data on …”
Barr Moses Apr 12, 2022 ▶ 13:15
Opinion
Moses: Traditional data catalogs are dead; long live automated data discovery
“We actually wrote a blog post not too long ago called data catalogs are dead. Long live data discovery.”
Barr Moses Apr 12, 2022 ▶ 18:26
Insight
Moses: Machines cannot infer semantic business logic without human input
“I don't think that a machine can actually infer that we can infer something without knowing that business knowledge. It's not possible, and that's also not what we attempt or attempting to do at Monte Carlo.”
Barr Moses Apr 12, 2022 ▶ 19:26
Prediction Not checkable as stated
Moses: Automation can resolve 80% of data downtime causes
“By introducing that level of automation, we can reduce our customer's team's work. Work from, you know 80% manual work to 20% manual work. So we can actually, you know, with the automation cover, cover 80% of reasons for why data downtime incidents happen and …”
Barr Moses Apr 12, 2022 ▶ 19:39
Insight
Moses: Generating actionable data quality alerts without alert fatigue is inherently difficult
“Making alerts meaningful making them ones that your team can actually act on is something that's very hard to do that we've invested a lot to do.”
Barr Moses Apr 12, 2022 ▶ 22:10
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 400 conversations transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.