Too many companies still treat IT spend like a necessary expense instead of a business lever.
That is where the EBITDA problem starts.
When technology spending is disconnected from the value chain, the result is usually the same: rising costs, more complexity, slow execution, and frustration on the front line. EBITDA stays flat, and leadership starts asking what IT is actually doing for the business.
I look at it differently.
Every company is here to make money. So technology spend has to connect to the value chain, the cost structure, and the risks that can hurt performance. If it does not, it is just expense.
I have spent more than 20 years working on that problem, including as CTO at Shoppers Drug Mart and Loblaw. Today, at Narrative Group, I help companies get the kind of technology leadership that links boardroom decisions to day-to-day operations.
The first question I ask is always the same: How do you actually make money?
That question matters because budgets are still going up. 96% of finance leaders expect to increase technology spending over the next five years. In a separate study, 91% of IT decision-makers plan to raise their budgets for 2025. So the money will move. The real issue is whether it improves EBITDA or disappears into maintenance, workarounds, and hype.
Start With How the Business Makes Money
When I begin with a company, I use a Financials First discovery assessment. I look at how it spends money on IT, how it deploys capital, and how technology shows up in the operation. Then I split the picture in two:
- the primary value chain
- the supporting functions that shape the overall cost structure
In the value chain, I am looking for productivity. Where is work slow? Where is it being touched twice? Where are people waiting on bad data or bad handoffs?
In the supporting functions, I am looking for leverage. How do we help the company do more without growing labor at the same rate?
Then I listen. Show me the day in the life. Show me what frustrates people. I also go on site, because clutter, bottlenecks, and manual side steps usually reveal where the undocumented process breaks down. That is often where you find technical architecture in conflict with the value chain.
If you are dealing with fragmented systems, start by drawing the flows. Who is talking to whom? How often? If your e-commerce platform does not match store inventory, I need to know whether orders are fulfilled from a store, a distribution centre, or both. I need to see how geography works, how inventory is set up by location, and how data moves when you change the pickup store.
Once that is mapped, the mess usually becomes clear.
This matters because most IT budgets are already under pressure. Industry data shows that roughly 70% of IT spend goes to run the business. If you do not understand where value is created, the budget gets swallowed by keeping the lights on. Follow the money first. Then you can decide what deserves investment.
Decide What Deserves Strategic Attention
One of the biggest mistakes I see is treating all IT spend as if it belongs in the same bucket.
It does not.
Vendors love to promise an all-singing, all-dancing tool. If the business model is unclear, that promise just gives you a more expensive problem.
I often say IT is effectively two departments. One side is foundational: networks, devices, security, collaboration tools, and core platforms. Those things need to be reliable, secure, and current.
The other side is strategic. That is where the team learns the business deeply, improves workflows, uses data better, and helps drive enterprise value.
So be ruthless. Categorize each function as one of three things:
- competitive advantage
- point of parity
- commodity
For the commodity pieces, use vanilla software. Stay current. Go mainstream, not leading edge and certainly not bleeding edge. Outsource the foundational work that does not make you special, and keep your best internal talent focused on the parts of the business that truly differentiate you.
I am also a big believer in outsourcing foundational commodity IT work to third parties. Desktop support and network support can be handled well by partners if the service levels are right.
This is where executives often fool themselves. They believe their process is unique when it really is not. In most businesses, the real differentiator sits in the people, how they serve customers, and how well they execute. The core systems around them usually need scale, security, and simplicity more than they need customization.
Cheap decisions here usually get expensive later. You save a bit now, then pay for rework when the company grows. I have seen companies cut license costs and quietly push the expense into labor. Then labor becomes the most expensive line on the P&L. I have also seen a company use an old tool called Mail Manager to push emails into SharePoint so it could avoid Microsoft license fees. Later, that shortcut got in the way of needed upgrades like Windows 11.
Freshworks found that companies waste about 20% of software budgets on unused tools and failed implementations. That is what low-cost thinking often hides.
Where CIOs Actually Move EBITDA
When a CIO aligns spend well, I expect to see four outcomes:
- more throughput
- better margin structure
- stronger reliability
- tighter capital discipline
That is where the EBITDA story lives.
Every business tends to have a productivity problem. That is one of the first assumptions I make.
So I measure work at the process level. How long does it take today? How much manual effort is involved? What is the rate of error or rework?
This matters because scale is usually limited by productivity long before it is limited by demand. If you want to double your customer base over the next 24 to 36 months, the business has to handle more volume without burning out the workforce. Standardization and automation are the first levers I pull.
Sometimes the gain is obvious. A task that takes 20 hours can be reduced to five. In a fixed-price business, that drops straight into margin. In a growth business, it changes the rate of hiring.
Usually the gain shows up in slower hiring, better deployment of talent, and more output from the same base. People move into higher-value work. The business gets more capacity without adding headcount at the old rate.
I saw this clearly at Shoppers Drug Mart. Project Symphony re-engineered the IT function into a portfolio-based matrix model, with specialized resources embedded where the work happened. Productivity improved by 40% year over year, measured against total invested capital. Better execution created more capacity from the same spend.
I have seen the same logic in automation work. In one case, a five-week e-commerce product setup cycle, which had taken one person three weeks of effort, was reduced to four hours. McKinsey has pointed to 30 to 200% ROI in many automation cases. That does not surprise me. When you remove tedious work, you create capacity the business can reuse.
Margin and Labor Economics
Productivity is one side of the equation. Margin is the other.
A strong CIO will craft an elegant operational business model. That is what innovation really is.
Self-checkout is a good example. The math looks simple: more lanes in less space, lower labor intensity, faster flow. But when the rubber hits the road, the economics depend on the user experience and the quality of the underlying pricing and product data. If customers need constant help, the labor goes right back in.
I lived both sides of that. At Shoppers Drug Mart, we piloted self-checkout in two stores. The user interface was simple, the pilot went well, and the rollout expanded to hundreds of stores. At Loblaw, the intervention rate was so high that we had to rewrite the front-end interface.
Same idea. Very different result.
I have also seen the margin story show up in back-office work. At Loblaw, more than 60 proofs of value were completed in 18 months, and more than 10 were scaled. Robotic process engineering across four functional areas produced more than $10 million in annualized savings. At Narrative Group, we advised a financial services technology provider and improved the P&L, adding more than 25 basis points to gross margin.
That is aligned technology spend. You can see it in the numbers.
Reliability, Uptime, and Risk
A lot of leaders still think of infrastructure as background noise.
Then it fails, and suddenly everyone sees how much work had been sitting on top of it.
Downtime hits EBITDA fast. Gartner has estimated the cost of mission-critical downtime at about $5,600 per minute. Your own number may be lower or higher, but the logic holds. When systems fail, labor stalls, service levels drop, and revenue gets harder to capture.
This is why I ask CFOs to quantify downtime in business terms. How many hours were lost last year? How many people were affected? What did that do to throughput?
Once you do that math, infrastructure stops looking like overhead and starts looking like operational drag.
At Narrative Group, we often build credibility by fixing the basics first. One client had six offices where people could not connect to the network when they moved between locations, and the meeting rooms barely worked. We replaced the core switches and network gear, refreshed old servers, put devices on a lifecycle plan, and cleaned up the meeting room experience. Three years later, the managing director told me everyone wanted the rest of the offices to look like the new one we had built.
Governance and Capital Discipline
The last lever is governance.
This one gets ignored until money is already being wasted.
A CIO who lives in firefighting mode usually misses the meetings where the real investment choices get made. Then the company ends up funding too many disconnected projects, with too little ownership and too little clarity. More spend does not solve that problem. It just makes the mess more expensive.
Formal portfolio discipline matters. One analysis found that companies with formal IT portfolio management achieved up to 50% higher ROI on technology investments. I have seen the same thing in practice. At Loblaw, I led the introduction of an IT strategy, a roadmap, and a governance model that aligned more than $300 million in capital spend to the enterprise strategy.
When we build a roadmap at Narrative Group, we organize it into three to five themes and balance quick wins with longer-range returns. Quick wins build credibility. The larger themes create the step change. Both matter.
You also need pilots. You need testing. You need leaders in the room making decisions with real information. During a major loyalty rollout, the solution worked in small numbers. Then we had to think about scale. What happens when millions of users hit the same endpoint in a short window on a Saturday morning? Performance testing delayed that rollout by three to four months. That was a good decision.
Sometimes the organization also needs to feel the cost of poor discipline. I have coached tactical IT leaders to stop rescuing every bad production decision. Let the problem surface in a controlled way. Once executives feel the disruption, they are finally ready to adopt proper change management and risk practices.
Governance affects enterprise value too. We helped a start-up insurance distributor define and communicate the use of funds within IT, put investment governance in place, and support a US$100 million raise. We also worked with Canada’s leading reverse mortgage lender on projects that improved execution and added to enterprise value. One project contributed to a valuation at a five-times multiple to one of Canada’s largest private equity funds.
AI Needs Adult Supervision
Right now, AI is the buzzword that creates the most confusion. There is real opportunity in it. There is also a lot of hype.
McKinsey reports that AI is consuming up to a third of change budgets while also adding to run costs. At the same time, industry research found that over 80% of companies report no discernible productivity gains from their AI work so far. That should make every CIO pause.
My view is simple. Start with the workflow. Look for long-running processes, weak data quality, repetitive decisions, and tedious work that people do not want to do anyway.
We are doing projects right now where AI and LLMs improve master data management across legacy systems and help clients delay massive ERP replacement costs. That is practical value.
I am wary when leadership wants to show off AI before the operating model is ready. Garbage in, garbage out is still true. The only difference now is that bad answers come back faster and sound more convincing.
You cannot just codify broken business processes. If the business rules are unclear, AI will automate confusion at speed. So find the boring foundational functions first. Document the process. Structure the inputs. Put governance around the work. Then bring in AI where it can improve quality, speed, and decision support.
Used properly, AI helps retool the workforce for higher-value work and gives the business more capacity. Otherwise, it becomes an expensive science experiment.
What the CFO Should See Every Quarter
If you want trust from the CFO and the board, talk in business language.
Talking in ambiguity will not help you.
The job is to give leaders enough information to make decisions and to own those decisions. What gets measured gets managed.
Every quarter, I want to see where the money went and what moved. I want to know how total IT spend is tracking against revenue and operating expense, and how the mix between IT labor and non-labor spend is changing. I want uptime, downtime, and the business cost of interruptions. I want named business processes with clear gains in speed, quality, or throughput.
I also want to see capacity created. If productivity improved, where did the labor go? Was hiring avoided? Were people redeployed into revenue, service, compliance, or customer-facing work? And I want the top cyber, compliance, and vendor risks called out plainly, along with progress across three to five strategic themes.
That is how we work at Narrative Group. We correlate IT spend against revenue and expenses, track service levels, and show leaders what contribution IT is making year over year. Once executives can see that movement, the conversation changes. They stop asking for shiny projects and start asking where the next pocket of leverage is.
If a proposed investment cannot be tied to throughput, margin, reliability, or scale, I challenge it.
Final Thought
I founded Narrative Group because I wanted to make top-tier enterprise technology advice accessible to small and mid-sized businesses, and because I want good things for Canadian companies competing on a global stage.
The path to better EBITDA is practical and measurable.
Start with one question: How do you actually make money?
Then follow the money, map the value chain, separate commodity work from real differentiation, fix the foundations, and measure what changes.
Do that well, and technology becomes a lever the leadership team can trust.
Frequently Asked Questions
What is the benchmark for IT spend as a percentage of revenue for mid-sized firms?
Mid-sized companies typically spend between 2 and 5% of revenue on IT. However, the exact percentage matters less than alignment. If that spend supports your primary value chain, it drives EBITDA. If it merely maintains legacy systems, it becomes an expensive drag on your margins.
How does shadow IT secretly erode our EBITDA and operational margins?
It often creates duplicated tools, hidden labor, inconsistent processes, and governance gaps. Over time, that can quietly increase cost and weaken scale.
How should IT KPIs be structured to protect or expand EBITDA?
Tie them to business outcomes. Track the effect of technology on throughput, downtime, labor leverage, service quality, and risk reduction, not just internal IT activity. Studies show organizations that align technology KPIs with business metrics grow 30 to 50% faster than peers.
With tech budgets rising, how do we make sure new investments generate EBITDA instead of overhead?
Force each initiative to show how it improves productivity, protects margin, reduces risk, or creates capacity. Without that discipline, rising spend just becomes a larger operating burden. Research shows companies waste 20% of software budgets on unused tools.
Should we prioritize AI or traditional process automation to maximize near-term EBITDA?
While over 80% of companies report zero productivity gains from AI, traditional Robotic Process Automation (RPA) consistently delivers 30 to 200% ROI. Usually, the better near-term move is the one with clearer process definition, faster payback, and stronger operating discipline. In many cases, that is traditional automation before AI.