Pull up the scorecard your leadership team reviews every week. Count the lines that are technology. There is a good chance the answer is zero, or one line that says something about project status. Sales has leads and bookings. Finance has cash and receivables. Operations has on-time delivery. People has turnover. Technology, which every one of those lines now runs on, appears in the room only when something breaks.
The weekly meeting is very good at the things that have numbers, and technology rarely arrives with one.
That is not neglect. It is a design gap. The weekly meeting is very good at the things that have numbers, and technology rarely arrives with one.
What the weekly meeting does well, and what it cannot see
If your company runs on EOS, you know the shape of the Level 10 Meeting. EOS Worldwide describes it as a weekly, 90-minute leadership meeting held at the same day and time every week. The Scorecard is a set of five to fifteen weekly numbers, each with an owner who reports whether it is on track. The Scorecard review takes about five minutes and involves no discussion. A number that is off track goes to the Issues List, and the bulk of the meeting, about an hour, goes to working the top issues in priority order. Other operating systems use different names for the pieces, and the lesson carries over wherever a weekly rhythm of numbers, owners, and issues exists.
The machine is excellent at what it can see. A sales number that dips gets named in the first five minutes and solved by the end of the hour. Now consider what happens to a risk that has no number. It cannot turn red. It cannot reach the Issues List until someone raises it as a headline, which usually means after the outage, after the breach, or after the invoice. Technology risk does not lose the argument in your meeting. It never gets a seat at the table.
Why technology stays off the page
Three reasons recur. The first is ownership. In most mid-market companies technology is run by a manager, a small team, or an outside provider, and none of them sits in the weekly leadership meeting. The second is the kind of number technology produces. Tickets closed, uptime against a vendor’s service level, and project percent complete describe activity, not exposure, and a CEO who has read them for a year has learned they rarely change a decision. The third is discomfort. Nobody wants to be handed a number they cannot interpret in the five minutes allotted.
It has to be one number, with one owner. It has to have a threshold that turns it red without a debate.
So the test for a technology number on the weekly page is strict. It has to be one number, with one owner. It has to have a threshold that turns it red without a debate. And it has to come from a system rather than from the opinion of the person responsible for it, a principle we laid out in The Control Group Is Gone. If a number passes those tests, it can sit between the sales line and the cash line without anyone needing a translator.
The five numbers
The first number is business-impacting downtime, in minutes this week. The goal is zero. Count only outages that stopped people from selling, shipping, billing, or serving customers, and let the owner mark the start and end from the monitoring system, not from memory. Cost is the reason this belongs on the page. ITIC’s 2024 Hourly Cost of Downtime survey, a web survey of more than 1,000 firms worldwide fielded from November 2023 through mid-March 2024, found that an hour of downtime costs more than $300,000 for over 90 percent of mid-size and large enterprises. Read that as a ceiling for your own thinking, not a forecast. Your company is probably smaller than the average respondent, which is why the better move is to compute your own number: revenue per operating hour, and what an hour of lost work costs in payroll.
The second number is critical security fixes overdue, as a count. The goal is zero. You set a deadline for critical patches and vulnerabilities, for example 14 days, and the scorecard shows how many are still open past it. The deadline is your policy, not a benchmark. The benchmark makes the case for a short one. Verizon’s 2026 Data Breach Investigations Report found vulnerability exploitation to be the leading way breaches begin, at 31 percent of breaches, and reported the median time to fully remediate a vulnerability rising from 32 days to 43. A count of overdue items is harder to flatter than a percentage, because one stale critical item cannot hide inside a 98 percent compliance rate.
The third number is return on technology investments. This is the line most companies cannot fill today, and that is the point of putting it on the page. PwC’s 29th Global CEO Survey, fielded from September 30 to November 10, 2025 across 4,454 CEOs, found that 56 percent reported neither revenue nor cost benefits from AI over the previous twelve months. The scorecard version is trailing-quarter return across your funded technology initiatives, measured against the hurdle you set before you funded them: benefit divided by total cost, with the cost including the people time and the ongoing run cost. Any initiative with no baseline and no per-unit measure does not get a flattering guess. It shows as unmeasured, and the count of unmeasured initiatives turns red on its own. The tests for what counts as a measurable benefit are in Proving AI EBITDA, and the way to build a comparison when everyone already uses the tool is in the Control Group piece.
The fourth number is an aggregate data quality score, with a goal above 90 percent. Choose the five to seven data sets the business cannot run without, typically customers, products or services, vendors, employees, and the financial chart of accounts. Score each one automatically on completeness, accuracy, timeliness, and uniqueness, and roll them into one weighted number. Two cautions keep it honest. The 90 percent line is the one we recommend starting with, not an industry standard, and an aggregate can hide a disaster, so show the weakest data set beside the aggregate on the same line. The reason to bother is the gap between feeling and measuring. A 2026 Precisely and Drexel LeBow survey of 505 data and analytics leaders at large companies, fielded in the second half of 2025, found that 67 percent reported high trust in the data they use for decisions, while 43 percent named data readiness as the most significant barrier to aligning AI with business objectives. Trust is a feeling. A score is a number you can watch move. Our Analytics Maturity Survey looks at where mid-market companies stand on the broader data, analytics, and AI ladder.
The fifth number is the share of known AI use cases that have a named owner and a risk tier. The goal is 100 percent. It is the inventory-and-accountability measure that The 9% Problem argues a board will eventually ask for, in a form short enough for a weekly page. It carries one built-in limit worth stating out loud: it measures only the use cases you know about. Verizon’s 2026 report found that 67 percent of users accessing AI services on corporate devices did so with non-corporate accounts, which is a good reason to assume your real list is longer than your known one. When you add a newly discovered use case, the percentage will fall. That is the system working, not failing, and it is why the owner’s job includes looking for what is not yet on the list. This is also where the logic of the Governance Freeway pays off: the number exists so that teams can move faster inside clearly marked lanes, not so that anyone can say no.
How to run it on Monday
Add the five as a technology block on the scorecard you already have. If your scorecard is close to its fifteen-line limit, replace the block with a single technology line that turns red whenever any of the five is red, and keep the detail on a second page. Either way, give each number an owner at your company. If an outside provider does the work, the owner is the person accountable for the provider, because a vendor can report a number but cannot own the consequence of it.
Set the thresholds in the first sitting, and take the first reading even if it is red. A red first week is information you did not have the week before, and it makes a better baseline than a clean number built on guesswork. Then follow the discipline the meeting already teaches. The scorecard review is for spotting, not solving. A red technology number goes to the Issues List like any other, and the hour of identifying, discussing, and solving is where the real work starts. At the end of the quarter, take the number furthest from its goal and make it the technology priority for the next ninety days.
What the scorecard will not tell you
Five numbers do not make a technology strategy. A green scorecard can sit comfortably on top of the wrong priorities, and a company can have perfect uptime for a platform it should have retired. The ROI line moves slowly and will usually lag the rest. Any number that becomes a target gets gamed, so keep a second source behind each one, and be suspicious of a line that has never been red.
It gives technology the one thing it has lacked in your weekly meeting: a way to turn red before something breaks.
What the scorecard does is smaller and more useful. It gives technology the one thing it has lacked in your weekly meeting: a way to turn red before something breaks. Our Vista Score rests on the same idea applied to readiness, a number that can be tracked instead of a reassurance that has to be believed. If you want help setting the first baselines and thresholds, an independent assessment is a good place to start.
Before your next Monday meeting, count the technology lines on your scorecard again. If the number is still zero, you already know what the first item on the Issues List should be.
