At our May event, Robert Kramer walked us through how the South Western City Schools District went about adjusting the boundaries for each school to account for both the shifting population and—the larger driver—the relocation, rebuilding, and new construction of various schools in the district.

Robert opened with a quote from another Robert – Robert Louis Stevenson:

I am told there are people who do not care for maps, and I find it hard to believe.

The author of this recap certainly agrees with Stevenson’s assessment (ESRI! GIS! .shp! FTW!), but, like any domain, it’s downright fascinating to pull back the covers and consider the space through a “data” lens. Little “Oh! Interesting!” moments occurred throughout the session:

• Catchments—do you know that word? I did not. It’s a term for each unique combination of [Elementary School] > [Middle School] > [High School] that exists within a school district. Put another way, if you’re ever chatting with someone about schooling in a town and ask, “Oh, so did your kids go to X middle school?” you’re in catchment territory. How many catchments are in South Western City Schools? 37! Could it be lower? Yes. But there are 16 elementary schools, 5 middle schools, and 4 high schools plus a variety of specialty schools. The theoretical minimum catchment count is 16, but that’s not practical because… yeesh! Complicated!
• An individual’s address is sensitive information, but addresses themselves are not!—the district rigorously maintains a database of every address in the district, and every address is assigned to a specific catchment. That’s totally public information (if you’re buying a house in the district, you would might want to look up which schools kids in the household would go to at various ages).
• Forecasting growth for middle schools and high schools has a leg up on forecasting growth for elementary schools—the district knows how many 6th graders are at each address in the district, which means they can reasonably predict how many 11th graders will be at each address in five years. That’s a little trickier for pre-K and kindergarten, though, because the school doesn’t have details on not-yet-enrolled, in utero, or not yet conceived kiddos!
• “Redistricting history” is an important consideration—when drawing new districts, one consideration needs to be whether an address was already redistricted in the recent past. It’s just…not cool… for a household to be getting whipsawed between schools every few years because they live in a boundary-ish location.
• School boundaries (and entire school districts!) can cut through apartment buildings (or individual buildings!)—that’s tricky! If you thought the phenomenon of “the kids on the next street go to a different high school than my kids” was odd, imagine that happening within a single apartment complex!
• Schools don’t have to be located inside the boundaries of the district they serve—this was almost a throwaway comment Robert made during the Q&A, but there are situations where if a student lived in the basement of a particular school (no students live in the basement of any of South Western’s schools; this isn’t Rudy), they wouldn’t attend that school. Obviously, this is a corner case and, in practice, doesn’t affect anyone.

Of course, all of the above points to the fact that there are a lot more considerations than just “near-term-perfectly-optimized-boundaries”. What Robert and his team developed in this case was an interactive (Tableau-based) planning tool whereby the team could selectively pick different areas on a map, remove them from one school and then add them to another school and see the impact across a multitude of relevant measures. Ultimately, they landed on three main options that were presented to staff and the public, and there were no strong opinions or debates and the final option was smoothly settled upon.*

As the final flourish on the presentation, Robert showed how, once the final set of boundaries was established, they were able to then automatically generate the SQL from within the BI tool to do a mass update the underlying database!

Of course, shifting populations and growth mean that this work is never truly done, but it certainly seemed like we were seeing an example of how data—properly cleaned and integrated, and then surfaced in a meaningful way—both met and will continue to meet the demands of that work!

We were also happy to have donated to the South-Western City Schools Educational Foundation on behalf of Robert for this talk.

Slides from the event are below:

* Okay… maybe there were some strong opinions and debates. Tradeoffs and competing priorities mean that’s going to happen, but it really did seem like the transparency of the process combined with the readily available impacts on a range of measures for each option, made this all go much more smoothly than its gone in other school districts going through a similar process recently (<cough> Dublin City Schools <cough>. I live there, but was blissfully unaware; it’s nice to have kids well beyond high school age).

And, of course, some pictures!

For our February event, we took a dive into a topic that affects anyone who is (or wants to) draw a salary: compensation. Alex Moore from Moore Cooperative walked us through the ins and outs of how companies figure out how much to offer their prospects and how that has to fit into the ongoing world of what they’re paying their current employees.

It’s a messy world of competing interests and priorities, and a misstep can quickly snowball: hire someone at a rate that is “too high” and then have them stick around for years with steady percent salary increases, and they can suddenly be compensated outside of the organization’s defined pay bands.

Of course, the pay bands are tough to maintain, too. Reliable market data has a limited shelf life, and figuring out the “right” compensation is more than just matching job titles. The same level and title in one industry may get compensated wildly differently in another industry (often because the role itself is quite different). The cost of living varies widely across geographic regions, too, so that has to be accounted for, but then what happens with remote workers who choose where to live (or who choose to move!)?

Did we mention that pay bands are nice idea, but they can be maddeningly challenging to put into place when an organization is working to maintain a strong and enduring workforce? According to one study, more than 20% of employees are paid outside their company’s official salary ranges!

Of course, compensation is more than just salary. Enter the “total compensation” discussion: health insurance plans vary widely when it comes to their coverage and cost, 401K matches can be anywhere from nonexistent to generous, paid time off can be flexible and expansive or stingy, and even in-office requirements can be Draconian or casual. Some of these aspects of compensation are negotiable, and both Alex and an attendee who is a full-time compensation analyst vigorously agreed that every offer should be negotiated!

Alex covered a number of additional aspects of this space:

• Varying regulations—country (although not the U.S.), state, and city-level requirements for pay transparency (the more you know: in Cleveland, employers with more than 15 employees must include salary ranges in job postings; of course, they can always try to take a page from Netflix and claim a salary range of “$150,000-900,000”)
• Varying efforts by companies to make their pay “fair”—from deep dive analysis of their comp program and processes to instituting remediation plans to committing to pay transparency
• Generational divides—one study showed that 89% of Gen Z employees are comfortable sharing their pay with their colleagues (which makes the Gen X author of this recap clutch his pearls)
• Gender pay gap—yep, it’s still a thing; it was at least closing there for a while until COVID came along and appears to have reversed that trend

The audience was engaged and had a lot of questions. It was hard to not get to some biggies, like the question about how we know information asymmetry generally contributes to inefficiency, so why don’t companies just go with full transparency as the norm‽  Well…it’s complicated. But it was fun to ponder with the group!

Slides from the event:

And some pictures!

Kicking off 2026 with a bang, we squeezed a healthy mix of long-timers and first-timers upstairs at COhatch Upper Arlington. We opened the event with a bit of a look back and forward on this meetup that has been running since 2008 (!) and followed up with presentation by one of our OG organizers, Tim Wilson, about KPIs.

Some of the highlights of Tim’s talk included:

• How KPIs are at the core of one key way that organizations use data: performance measurement
• How performance measurement—when done well—is the construction of a metaphorical time machine: we establish clear, outcome-oriented KPIs as a way to align on our expectations for what results we will achieve with a campaign, project, or initiative; that then allows us to look at results (in the future) and travel back in time (metaphorically) to objectively compare those results to our expectations.
• This is simple, right? But not easy?
• Can AI do that? Can we just ask AI. “How did my campaign perform?” We can, but the best response it will give will look like the response that a pretty lousy analyst would provide to the question: a puking of data with some arbitrary comparisons to other data that it can access. So, no. We can’t just ask AI. Performance measurement is about humans aligning on expectations for business outcomes.
• What does work for this? Asking two magic questions: 1) What are we trying to achieve (with this effort)? and 2) How will we know if we’ve done that? That second question is a two-parter: it requires identifying one or more direct or proxy measures (KPIs) and targets for each of those measures.
• Business teams (marketers are particularly guilty of this) loathe setting targets. It freaks them out. They have a lot of good-sounding excuses for why they can’t set targets.
• But they’re wrong. No targets. No time machine. Ineffective performance measurement.
• A (Mini) Wisdom of the Crowds approach is a great way to set targets, though, and everyone gets on board quickly: just have everyone come up with a target (their expectation) independently (since everyone’s got the assignment, no one feels individually exposed) and then share what proposed targets they came up with. This will always spark a thoughtful discussion and a quick alignment on what the target (or target range) should be.
• AI also comes into this process when we talk about AI initiatives. They, too, can have their performance measured with those two magic questions—what business outcome is the effort trying to achieve, and how will that be measured?

Lots of good stuff. Tim even dressed up as the AI cartoon that was sprinkled throughout the presentation. And, he measured the performance of the session itself in real-time by polling the audience at the very end, which was also an opportunity for him to give away a few signed copies of his book, Analytics the Right Way, and a signed copy of John Lovett‘s book, The *NEW* Big Book of KPIs. All of that was really an excuse for him to create an R script that he’d have to run on the fly during the presentation. Despite creating the perfect opportunity for this scripting to fail, Tim’s prayers to the gods of R paid off and everything worked.

The slides he used (including the results of the measurement of the session itself! Spoiler: he handily exceeded the target for the two KPIs he’d established) are available below:

Did you know that Data and Analytics Wednesdays were initially started (not in Columbus) as pure socializing/networking events? From the get-go, we’ve always included an educational component in ours because, well, you know, midwestern-purposefulness or something.

We break that format once a year in December with a content-free event. And, in 2025, we tried to break it a second time by having our July event be a group outing to a Columbus Clippers baseball game.

Alas! We failed to go entirely content-free because baseball, after all, is the OG analytics-oriented sport! Bill James! Sabermetrics! Moooonnnnneeeeyyyyybbbbbaaaaalllll!

So we had a little bit of content. Most of the attendees arrived early so that a few members of the Clippers scoring team could pop down to our seats for a little Q&A that was as fascinating as it was informal!

And then we watched the game!

The discussions in our seats (alas!) drastically outperformed the Clippers’ performance on the field. The Louisville Bats were up 7-3 halfway through the 4th inning, and the score remained there for the duration of the contest.

But, with each attendee armed with some Clippers Cash, a koozy, a pass to the Tansky Club behind home plate, and great weather, an excellent time was had by all!

Our May 2025 event was all about communication. Specifically, it was about all the various flavors (or “genres”) of speaking publicly, be that on a conference stage, in a conference room to a group of stakeholders, in a client’s office for a high-stakes pitch, or to a camera on a video call where the participants may or may not have their cameras turned on.

Ruth Milligan and Acacia Duncan, two thirds of the author trio for The Motivated Speaker: Six Principles to Unlock Your Communication Potential walked a 50-strong audience of engaged attendees through an introspective and interactive exercise in identifying their “shaded (ineffective) habits” when it comes to public speaking.

What is a “shaded habit?” It’s something (or multiple things) that every individual has picked up over the course of their lives that feels natural and comfortable even as it gets in the way of effectively communicating.

The bad news? Everyone has them.

The good news? No one was born with any of them, so whatever those habits are for an individual, they can be identified and unlearned (or, at least, sufficiently mitigated).

Ruth and Acacia opened the session with some ripped-from-the-headlines ripped-from-their-clients-with-identifying-details-removed of communication failures and led a discussion with the attendees as to the root causes of those failures. From there, they prompted everyone to think about their own rhetorical style and what they could identify as their shaded habits. Attendees jotted their thoughts on Post-it notes that Ruth (and Tim) collected and grouped for review and discussion:

The most commonly identified habit? Overuse of “filler words”: “um”, “like”, “you know”. How to address it? Breathe! And shorter sentences. With pauses that give the period its due. [artistic license intentionally taken on the preceding sentence fragments. To make a point. Just did it again.]

Other types of shaded habits that came up included: rambling, talking too fast, not minding the clock, going into too much detail, not thinking sufficiently about the audience’s needs (what questions they want answered rather than what information the speaker wants to share), and more!

Some of the shaded habits were diagnosed as being different forms of stress responses, of which there are fundamentally four distinct flavors: fight, flight, freeze, and fawn. The tricky thing about stress responses is that they’re not going to just go away. They’re going to happen. But, by recognizing what our default flavor of stress response is, we can prepare for how to deal with it, be it by lifting heavy weights just before speaking (for real…!) or “grounding” ourselves (anchor feet to the floor, hands palm down on the table if sitting) or repeating a mantra (“This too shall pass” may work, but it can be whatever works for you!).

Following the exercise and discussion, we had a drawing to give away five copies of The Motivated Speaker to lucky audience members and then had a book signing (20% of the proceeds from the book sales went to Sanctuary Night).

As the emcee noted at the start of the meetup, our goal is for every event is for attendees to take something away that they can put into action within a week, and our May event absolutely delivered on that front!

Additional pictures from the meetup are below:

We mixed things up a bit for our September event, both with location and format. Given the topic, we hosted the event at Denison Edge, which is a really cool venue!

The topic was inspired by the experiences of a rising senior at Kenyon College that, despite excellent qualifications and impeccable due diligence, barely managed to land an analytics internship in the summer of 2024. Some relevant details of that internship:

• The company that hired him was a small agency that had not really thought about having an intern
• Through a string of improbable but fortunate events, they hired him for the summer
• The student had a great experience, and the agency found that he added real value to their work
• Things went so well that the company kept him on for ~10 hours/week once he returned to school in the fall

That’s the happiest of endings, sure, but the CbusDAW organizers were struck that it was almost certain that this specific tale represented countless simlar stories that never came to pass. And that’s a miss.

Consider:

• Companies of all sizes (including small ones!) have data at their disposal that is underutilized due to a lack of resources
• College students today—across all types of programs—are developing valuable skills in programming, statistics, and analytics in the classroom
• Academic programs recognize the importance of their students getting hands-on, real-world experience, and there are any number of resources in place to support getting them that experience

We brought together four panelists from central Ohio-based higher education to have a discussion about getting all of those factors to work together to create more win-win situations. The panelists:

Matt Miller Denison University	Nimet Alpay Franklin University
Tom Metzger The Ohio State University	Kristen Astorian Ohio Wesleyan University

While the initial idea for the panel was “internships,” the panelists made it clear that internships are simply one way for students to get real-world experience while delivering value to organizations. Many data and analytics programs—both undergraduate and graduate level—require a capstone project that works with an organization and their data to deliver value (and capstone projects have the benefit of having instructor oversight and coaching).

Some keys to making an internship successful:

• The project should be meaningful—using interns to work on projects that are menial doesn’t benefit the intern or the organization that hired them
• The project should be manageable—dropping an intern into a monstrously complex data environment with overly ambitious ideas for what they will be able to deliver in a finite period of time is setting them up for failure
• The intern should have a primary point of contact for oversight—this should be someone who actually wants to take on the work. They’re playing the role of a guide, mentor, and manager all at once.
• Consider pairing the intern with someone deeply knowledgeable of the data itself—it can take months to ramp up on the intricacies of many organizations data environments. While students do need to get exposure to the messiness of real-world data and the often-daunting level of effort to “clean it up” as part of a project, it can be useful to have someone who knows the ins and outs of the various tables assist them in getting queries written.

There are also a surprising number of programs (if only the moderator of the panel was not also the author of this post—something of a hindrance to note-taking!) that provide support to companies who are open to taking on interns (or to working with students on capstone or other projects):

• The career centers at most universities have staff who are deeply familiar both with there students and what it takes to scope work and provide support in order to make student work productive and impactful
• Through various programs (a range of funding sources), companies can actually have interns’ pay subsidized (partly or fully)! The career centers at any school can point interested companies to resources for that.

It was very clear that, once an organization tries out tapping into student talent, they consistently extend and expand their programs over time. Have you given that a thought? Reach out to one or more of the people above to find out more!

For this month’s event, Matt Gershoff, CEO of Conductrics, traveled from the land of triple-digit heat on the regular (Austin) to the land of pleasant-temps-even-in-August (Columbus) to share some thoughts and examples about being intentional when it comes to data collection. If you attended the meetup, then you’re here because we promised that we’d post some useful supplemental information, and we’re going to flip the script of a normal recap post by putting those up front:

• [20-Minute Video] Matt’s talk at PEPR ’24—a Venn diagram with the talk he gave at DAW has something like a 63% overlap, although his DAW talk is a larger circle, as there was additional material! But, since we don’t record DAW talks, the PEPR talk is a good one to share with a colleague who is kicking themselves for not attending DAW.
• [<2-Minute Video] Matt talking about intentionality—not even remotely by design, this was a piece of an interview that another one of the DAW sponsors, Piwik PRO, did with Matt. Useful and thoughtful stuff.
• [5-Page PDF] Privacy by Design: The 7 Foundational Principles—a very worthwhile read; Matt focused primarily on Principle #2, but you’ll never believe what Principle #4 and #7 are! (Seriously, if you give it a real read, it will blow your mind a little bit; it’s almost three decades old and was an underpinning of GDPR!)
• Matt will also be on an upcoming (as of this writing) episode of the Analytics Power Hour podcast, so, if “audio only” is your or a colleague’s jam, smash the subscribe button there.

Matt’s presentation—with annotations added to make it an upgrade from “just the slides”—is included at the end of this post, but a few of the highlights from his presentation were:

• “Just Enough” vs. “Just in Case” Data Collection—Matt made a stronnnnnng case that the industry bias is for the latter, while “privacy by default” demands the former. “Just Enough” data means aligning to a specific and explicit task or objective and then collecting as little data as needed to complete the task. “Just in Case” is a “maximize optionality” play—hoovering up as much data as possible at as granular a level as possible so that there are as many possible “options” for doing “stuff” with it in the future. We are so wired to the latter that it’s uncomfortable to recognize why that Is. Not. Good.
• This doesn’t mean there are no cases where high granularity / high cardinality data is warranted—throughout the talk, Matt was clear that he was not speaking in any absolutes (unless we count as an absolute that, “all data collection should be performed with intentionality”).
• Many types of A/B tests, both univariate and multivariate, can be statistically evaluated without recording data at the individual user level—if you’re like the author of this recap, you’ve always operated under the assumption that A/B tests require capturing each user’s session, including which variant they were assigned to, maybe some other meta data about them (what level of a loyalty program they belong to, for instance, for “deeper dive analysis”), whether or not they converted, and maybe the amount of their purchase. 10,000 visitors in the test? That’s 10,000 rows of data! What Matt demonstrated was, um, no. That’s incorrect thinking. By using equivalence classes, some understanding of the maths underlying the statistical tests of interest (t-test, OLS regression, and more), it’s possible to simply capture/increment aggregated counts (visitor count, sum of sales, sum of the squares of sales) and perform the exact same statistical tests in a way that is: computationally less intensive, data storage much less intensive, and aligned with privacy by design principle #2: privacy by default (and privacy by design principles #3 and #4 and #7). Matt outlined a lot of this in this blog post (although he has since extended his research and thinking on the subject… and it continues to hold up!)
• There are different techniques and concepts that are good to be familiar with when embracing privacy by design—K-anonymity, differential privacy, global vs. local privacy, and more! The key with all of them is that they’re best employed when approaching them as privacy by design rather than privacy-tacked-on-later-to-maintain-regulatory-compliance.

A lot of ground was covered with pretty lively audience engagement and more than a few laughs!

The annotated slides:

And, as always, a few pictures to capture the atmosphere:

At our May 2024 event, Nick Woo from AlignAI shared a thoughtful and pragmatic perspective about how to approach figuring out what use cases are (and are not!) appropriate for AI. The turnout for the meetup was strong, and the discussion was lively!

Nick started off with a handy definition of machine learning:

“Machine Learning is an approach to learn complex patterns from existing data to make predictions on new data.”

Oh. Sure. Seems simple enough, right? But that doesn’t include generative AI, does it? As a matter of fact, it does:

• The existing data is what was used to train the model
• The new data is the prompt that is provided to the model (!)
• The response to the prompt is really a prediction when the model processes that new data (!!!)

Nick also outlined the anatomy of an AI use case:

1. Business Problem
2. Data
3. Training
4. Model
5. Accuracy Metrics
6. UX/UI

Which step is the most common stumbling block for organizations’ proposed use cases? The “Data” one—there needs to be sufficiently scaled, cleansed, and complete data to actually develop a model that is useful. Oh, and then that model will likely need to be refreshed and refined with new data over time.

The most neglected step in the planning of an AI project? The last step: actually thinking through what the user experience should ultimately be when the model is put into production!

Nick was quick to point out that it is easy to treat AI as a hammer and then seeing all the world as a nail. If there is a simpler, cheaper, equally effective way to address a particular business problem, then addressing it with AI probably doesn’t make sense! He also acknowledged (as did several audience members) that we’re currently at a point where there are executives who truly do just want to be able to say, “We use AI,” which means some projects can be a bit misguided. This phase shall pass, we assume!

Another discussion that cropped up was measuring the ROI of an AI use case. Nick noted that this can be shaky ground:

• AI technology platforms pushing to measure impact simply based on the adoption of the technology (rather than quantifying actual business impact)
• Minimal use of techniques like controlled experimentation to quantify the impact (there is simply too much excitement currently to create interest in withholding the magic from a control group in a disciplined way)
• The ROI of an AI project can be thought of as “the ROI of an OPEX project”—organizations that are disciplined about measuring the impact of non-AI OPEX projects should be pretty good about quantifying the impact of their investments; it’s just another tool in their toolkit, so the measurement mindset can be the same

And… there was more, including an example scoring matrix for prioritizing use cases across multiple criteria!

A recap post and the slides really can’t do the evening justice, but it’s better than nothing. The recap was above. The slides are right here:

And some pics from the evening:

We tried something a little different in this month’s DAW. We actually tried two things that were a little different in this event.

What we intended to be different was that we were going to have a panel of experts who would field a bunch of questions from the audience, capture them on a whiteboard, and then talk through them. Ultimately, we did that—not exactly as it had been drawn up (so to speak), but it worked out.

The unintended difference in the event was to see how many things could go wrong and still have us pull off a successful and engaging meetup. Speculation was that the questions and answers were going to be so good that our robot overlords became concerned and flexed their AI capabilities to undermine the meetup. To wit:

• On Monday, one of the three intended panelists pulled out of the event. No problem, Brian Sampsel was hastily recruited and graciously accepted the last-minute invitation.
• On Wednesday morning at 4:00 AM, one of the other panelists went into labor. Did she take the time to email us that she had become unavailable? Yes. Yes she did. Katie Schafer is a machine in her own right (as our other panelist, Pete Gordon, had already noted several days earlier). But, no problem. We could do this with two panelists. What else ya’ got to throw at us, HAL? Well…
• Weather anyone? The venue and the surrounding area had a tornado watch issued late afternoon, and the venue—Rev1—was squarely inside the tornado watch area. The tornado watch lasted until 7:00 PM (the event started at 6:30 PM). There was rain. There was wind. There was hail for Pete’s sake!

Apparently, though, analytics types take their cues from the USPS. Or have poor judgment. Or some combination? We wound up with a great turnout, with lots of good pre-talk discussion over pizza and beer:

Conveniently, the event is in the interior of the building! #tornadosafety

The discussion itself covered a wide range of topics—skewing heavily towards AI and less to date science (data science is involved in AI, of course, so it was still there):

A Range of Topics to Discuss

There is no deck to share, no recording, and this attendee didn’t take scrupulous notes, so we’ll go with a smattering of the discussion that could be retrieved from his brain the following day:

• When will AGI (artificial general intelligence) be achieved? Pete’s estimate (which seemed serious) was: 2033. But, he also noted that AlphaGo’s infamous Move 37 (in 2016) was a glimpse into that future.
• To RAG or not to RAG? Well… that’s a hot topic. It depends.
• Poisoning of training data? Why, and what are the ramifications? It sounds bad, but it’s got it’s uses—see Nightshade.
• Should newly minted software engineers be worried about AI making their jobs obsolete? No. Full stop. They’ll have some powerful new tools to employ, but their jobs aren’t going anywhere.
• What about marketing analysts? Will AI take their jobs? This prompted quite a bit of discussion. Brian made the point that AI can do some pretty impressive exploratory data analysis (EDA), which is definitely useful! One attendee asked if he could see getting to a point where you could tell an AI-based tool what your KPIs were, and it could then just analyze the campaign. The answer was, “Yeah… but a human still needs to set appropriate KPIs!” Even MMM came up—is that AI, or is that just… sophisticated linear regression (statistics). Kinda’ more the latter, but “AI” gets slapped on it for branding purposes and we get excited!

And, of course, lots and lots more! Some pics from the event:

This month’s meetup was our first since we were rechristened as “Columbus Data & Analytics Wednesday.” In an unintentional twist, the topic for the event was centered around the speaker’s contention that we (the broad, collective “we”) devote too much of our time and energy to the collection and management of data, and not enough effort to actually putting that data to productive and impactful use within our organizations.

Tim started out by calling out that, if we consider any task that has any relationship to data as “data work,” then we can further categorize each of those tasks into one of two buckets:

• Data Collection and Management
• Data Usage

He noted that there is no inherent business value in the collection and management data. There is only the potential for value. To realize that value requires putting the data to meaningful, applicable business use.

All too often, data workers get so caught up in data collection and management tasks, though, that they start to believe that there is inherent business value in those tasks alone. Tim pointed to three reasons for this happening:

• Technology vendors tend to have business models that are high fixed cost and low variable cost, which means they’re incentivized to drive aggressive customer growth. This results in heavy investments in marketing and sales organizations that wind up distilling down their messaging to, “Buy our technology and you will realize business value.” And they spend a lot of time and money promoting that message.
• Consultants have the opposite business model—low fixed costs and high variable costs—which means they grow profitably by selling engagements that use repeatable processes that can tap into a scalable workforce. That pulls them to “technology implementation” work over “deeply engage with the businesses of their clients and all of the complexity therein.” So, they wind up promoting a similar message: “Buy (our partners’) technology, let us help you implement it, and you will then realize business value.
• Human nature within organizations drives us to do tangible “things”—adding new data sources, cleaning up data quality issues, building or augmenting dashboards, etc. This leads us to telling ourselves that these tactical activities, which skew heavily towards data collection and management, bring value to the business in and of themselves.

According to Tim, recognizing and pushing back against this mindset means embracing the messiness and hard work required to actually use data productively. He proposed that organizations need to put the same level of rigor around their data usage processes as they put around their processes for collecting and maintaining data. As an example, he outlined a framework he uses (but was clear that this wasn’t “the only” framework that’s valid for data usage) that pointed to three distinct ways data can be used to provide value:

• Performance measurement—objectively and quantitatively answering the question: “Where are we today relative to where we expected to be today at some point in the past?” He described using “two magic questions” for this: 1) What are we trying to achieve, and 2) How will we know if we’ve done that?
• Hypothesis validation—this is all about improving decision making by reducing uncertainty when determining what to do going forward. For this, he described a 3-part fill-in-the-blank technique: “We believe [some idea] because  [some evidence or observation]. If we are right, we will [take some action].”
• Operational enablement—data when it is actually part of an automated or mostly automated process (for instance, ordering shoes online generates data that is used by the order fulfillment process). He went on to say that every generative AI use case he’s seen put forth falls into operational enablement.

He ended by imploring the crowd to look at the work they and their colleagues do day in and day out through a “data collection & management” vs. “data usage” lens and consider working to shift the balance of their efforts towards the latter!

The slides are available below, as well as at :

And, of course, some pics from the event, which had a large and lively showing!