FoundersOperators, strategy & lessons
Dave Kellogg argues executive teams should review one shared table of numbers rather than separate dashboards
Former Host Analytics CEO Dave Kellogg argues that executives with their own dashboards still aren't discussing the business together. He says better decisions on segments, events and hiring come once metric definitions are settled and the team reviews one carefully chosen table together.
From the Kellblog article of November 16, 2026 and cited sources: Kellogg's central claim is that separate dashboards for each executive don't lead to a real conversation about the business. [1]
SOURCES[1] The Awesome Power of Teams Sitting Around a Table of… kellblog.com
AI-assisted, not human-edited. Methodology Corrections Who owns this site
- What happened
- Kellogg's central claim is that separate dashboards for each executive don't lead to a real conversation about the business. [1]
- Original source
- Kellblog article, independent perspective
- Brief by
- AT Pulse research desk AI-assisted; publisher and contact on the desk page
- When
- November 16, 2026 article published
- Covered
- Nov 17 edition
- Evidence
- 7 sources across 4 sites
- 1The Awesome Power of Teams Sitting Around a Table of…kellblog.com
- 2Hidden profile – Wikipediaen.wikipedia.org
- 3Balderton Capital welcomes Dave Kellogg (press release)balderton.com
- 4Talking About the Numbers vs. Talking About the Business –…kellblog.com
- 5Shared information bias – KU Open Textopentext.ku.edu
- 6About Dave Kellogg – Kellblogkellblog.com
Reach: not measured yet

Kellogg's central claim is that separate dashboards for each executive don't lead to a real conversation about the business. That conversation happens when the leadership team sits together around one curated set of numbers that everyone understands. [1]
His earlier post gives the method. First agree on what metrics such as MQLs and forecasts mean. Then have ops fill in a shared sheet every week, show each figure beside prior quarters and years, review it together at the staff meeting, and settle disputes over definitions outside the meeting. [4]
Research on group decisions fits part of the premise but points both ways. Groups tend to miss facts that only some members hold, which a shared table helps with. But groups also over-weight information everyone already has, especially on judgment calls like where to invest. That suggests the table should invite challenge, not just agreement. [2] [5]
Studies show groups often overlook facts held by only some members. That partly fits Kellogg's premise but doesn't test his format.
Read the full assessment
A reasonable inference for leaders and builders of analytics tools: agreed definitions and joint review may matter more than adding dashboards.
Executive brief
The most consequential point is easy to overlook: Dave Kellogg argues that a team of capable executives, each with their own dashboard, is still not having a real conversation about the business. That only happens when they sit together around one carefully chosen table of numbers that they all understand (Kellblog, 2026-11-16). This is management commentary, not an AI product launch or a funding round. It contains no model, no metrics and no investment figures. Most of the post is behind a paywall, so the reviewed text covers only its premise. Decades of research on how groups share information fit the premise in part, but they do not test it directly.
What changed and event timeline
- Context
1985: The "hidden profile" problem is identified
Stasser and Titus showed that groups spend most of their discussion on information everyone already holds. As a result, they often miss the best option when the key facts are spread across different members ().
Meta-analysis on team information sharing
Mesmer-Magnus and DeChurch found that information sharing matters most for tasks with a correct answer, and less for tasks where the group only needs to agree ().
Kellogg joins Balderton
He became Balderton Capital's first Executive in Residence, advising its portfolio companies on SaaS metrics, strategy and expansion ().
"Talk about the business, not the numbers."
In an earlier post, Kellogg described an ops meeting that lost about an hour arguing over definitions: what counts as an MQL (marketing-qualified lead), how to reconcile forecasts, and how to credit deals ().
The shared table of numbers
The new post argues that once the definitions are settled, a group discussing the business together around one table makes better decisions on segments, events and hiring ().
Capabilities and access
- There is no software, model or version involved. The "capability" is a meeting practice.
- Only the opening of the 2026 post is public. The rest is for paying subscribers (Kellblog).
- The 2024 companion post is free and gives the concrete method: ops fills in a shared Google Sheet every Sunday, the numbers are shown next to prior quarters and prior years, the staff meeting reviews them together on Monday, and arguments about definitions are taken offline (Kellblog).
Read the full section
- There is no software, model or version involved. The "capability" is a meeting practice.
- Only the opening of the 2026 post is public. The rest is for paying subscribers (Kellblog).
- The 2024 companion post is free and gives the concrete method: ops fills in a shared Google Sheet every Sunday, the numbers are shown next to prior quarters and prior years, the staff meeting reviews them together on Monday, and arguments about definitions are taken offline (Kellblog).
Technical analysis for researchers and developers
The posts describe no architecture or evaluation. The implementation points they do document: For builders of analytics tools or AI summaries for executives, a reasonable inference is that consistent definitions and a shared view matter more than letting each person explore on their own. The sources do not test this.
Read the full section
The posts describe no architecture or evaluation. The implementation points they do document:
- Definitions come first. The team has to agree on what metrics like MQL and forecast mean before the numbers can support decisions (Kellblog 2024).
- One curated view. Everyone looks at the same table, not separate dashboards (Kellblog 2026).
- Comparison built in. Each figure appears beside its history (Kellblog 2024).
For builders of analytics tools or AI summaries for executives, a reasonable inference is that consistent definitions and a shared view matter more than letting each person explore on their own. The sources do not test this.
Claims and evidence
- Claim: Discussing the business with shared numbers beats discussing it without them. Evidence: This is the author's experience-based opinion (Kellblog 2026).
- Claim: Many companies think they are discussing the business when they are still arguing about the numbers.
- Independent support (indirect): Research on group decisions shows that groups lean toward information everyone already shares (Shared information bias, KU Open Text).
Read the full section
- Claim: Discussing the business with shared numbers beats discussing it without them. Evidence: This is the author's experience-based opinion (Kellblog 2026). The reviewed sources contain no controlled study of the format.
- Claim: Many companies think they are discussing the business when they are still arguing about the numbers. Evidence: This is also the author's view, illustrated with one anecdotal meeting (Kellblog 2024).
- Independent support (indirect): Research on group decisions shows that groups lean toward information everyone already shares (Shared information bias, KU Open Text). The reviewed sources do not show independent corroboration of Kellogg's specific format.
Context and prior work
- This is the third post in a series.
- Kellogg was previously CEO of Host Analytics (2012–2018) and MarkLogic, ran Salesforce's Service Cloud business, and was CMO of Business Objects (About Dave Kellogg).
- The hidden-profile research tradition is the closest academic counterpart (Hidden profile – Wikipedia).
Read the full section
- This is the third post in a series. The others are the 2024 post on numbers versus business and a post called "The Market Segment Analysis Chart," whose date could not be confirmed (Kellblog 2026).
- Kellogg was previously CEO of Host Analytics (2012–2018) and MarkLogic, ran Salesforce's Service Cloud business, and was CMO of Business Objects (About Dave Kellogg).
- The hidden-profile research tradition is the closest academic counterpart (Hidden profile – Wikipedia).
Limitations, safety and contested findings
- Most of the argument sits behind the paywall, so the reasoning the post promises (why the shared conversation beats individual dashboards) could not be checked.
- The research cuts both ways:
- Showing everyone the same table removes information gaps, which helps.
Read the full section
- Most of the argument sits behind the paywall, so the reasoning the post promises (why the shared conversation beats individual dashboards) could not be checked.
- The research cuts both ways:
- Showing everyone the same table removes information gaps, which helps.
- It could also strengthen the bias toward information everyone already holds. That bias is strongest on ambiguous, judgment-heavy tasks where the goal is agreement (KU Open Text).
- Questions like "which segment should we invest in?" are judgment tasks of that kind.
- Research links dissent to better decisions when key information is unevenly held (Hidden profile – Wikipedia). That suggests the table should invite challenge, not just agreement.
Business and practitioner implications
- Leaders: Budget a few quarters, not weeks, to agree on metric definitions before expecting strategy debates to be productive (Kellblog 2024).
- Ops and data teams: Each week, publish one table everyone has agreed on, with historical comparisons.
- Builders of AI and BI tools: Shared definitions and group review may matter more than more dashboards.
Read the full section
- Leaders: Budget a few quarters, not weeks, to agree on metric definitions before expecting strategy debates to be productive (Kellblog 2024).
- Ops and data teams: Each week, publish one table everyone has agreed on, with historical comparisons.
- Builders of AI and BI tools: Shared definitions and group review may matter more than more dashboards. This is an inference, not a tested result.
- Venture desk: There is no funding round, valuation or investor in this story. Kellogg advises Balderton's portfolio, but the post does not name any companies (Balderton).