How to Build an IT Roadmap When You Don’t Have a CIO

Last updated: October 8, 2026

By Colin Boothe, CIO at Foretopia

If you do not have a CIO, build the IT roadmap backward from dates you do not control, which means renewals, end-of-support deadlines, acquisitions, and go-lives, not from a list of things you want. Four quarters, one page. Every line names the decision the date forces and who does the work.

One date that belongs on the roadmap and rarely makes it there is the shutdown date for whatever you just replaced. We broke down deciding what happens to the system you replaced, including the consumer inventory, the retention rules, and who should own the kill date.

The roadmap was a Word document called ROADMAP_FINAL_v4.docx.

Nine initiatives in it. A data warehouse. “Evaluate AI opportunities.” Two different single-sign-on projects, written by two different people, eighteen months apart. Not one date anywhere in the file, except the one in the filename, and that version had been sitting in a SharePoint folder marked FINAL since some point in 2023 that nobody present could pin down for me.

The CFO owned IT at this place, a $180M specialty distributor, which is normal and which nobody says out loud. She was good at her job. She also had a quarter-close, an audit, a bank covenant with a fixed-charge coverage ratio she checked more often than she admitted, and the IT roadmap, which was the fourth thing on a list of three. So it sat.

Then two things landed in the same month. Their EDI provider sent a renewal with a 22% increase and a 60-day notice window, and in the same week somebody finally said out loud that the company they had just acquired ran a different ERP on a different fiscal calendar. Suddenly there was a roadmap. It just wrote itself in a panic instead of on purpose.

That is the whole problem. Not strategy. Timing.

Quick disclosure, because you should want one. I run a consulting group that sells fractional tech leadership for a living, so a page arguing that you should rent a part-time CIO instead of hiring one is about as neutral as a barber telling you it is time for a haircut. You are also reading it on a staffing firm’s site, and KORE1’s IT staffing services desk gets paid when the roadmap turns into open reqs. Both true. The dates below are still the dates.

What an IT Roadmap Actually Is

An IT roadmap is a dated plan that maps your technology decisions onto the next four to eight quarters, showing what gets decided, what gets built, and who does the work. It is not a wish list, and it is not an architecture diagram. Its job is to force a decision before a date forces it for you.

Most mid-market roadmaps fail the second half of that definition. They describe a destination without naming a single decision or a single person. Which makes them a mood board.

Here is the test I use. Pick any line on the roadmap. Can you say, out loud, what decision it forces, by when, and who is accountable? If you cannot do that for a line, delete the line. I am being slightly harsh. Only slightly. But a nine-item roadmap where you can only answer that question for three of the items is really a three-item roadmap carrying six items of decoration, and the decoration is what makes the whole page feel too heavy to pick up.

Start With the Dates You Do Not Control

Nobody wants to start here. It is clerical, it feels beneath the exercise, and there is no version of this task that makes anybody look visionary in a steering meeting. Do it anyway. It is also the only part of a roadmap that cannot be faked, which is probably why it gets skipped.

What you are collecting is every date somebody else already set for you. Order forms, not invoices. Published support lifecycles. Lease ends. The TSA exit date on whatever you bought in March. One client’s single most expensive surprise turned out to be a storage array lease that auto-renewed every October, and three different people each assumed one of the other two was tracking it.

These are not priorities. They are gravity.

The end-of-support row is where mid-market companies get genuinely surprised, so here is a live one. Microsoft ends Dynamics GP support for product enhancements, tax updates, and technical support on December 31, 2029, with security patches available at Microsoft’s discretion until April 30, 2031. GP also closed to new customers along the way, perpetual licenses in April 2025 and subscriptions in April 2026.

Read that again if you are on GP, because the dates lie to you in a specific way. 2029 sounds like somebody else’s problem. It is not. A finance-system replacement at a $150M company is an 18-month exercise once you count selection, data cleanup, a build, parallel close, and the quarter where you do not trust either system. Back 18 months off a 2029 date and you are scoping in 2027, which means budgeting in 2026, which means now.

Also, and almost nobody checks this, older GP installs still governed by the fixed lifecycle are already past the wire regardless of what the 2029 headline says. Extended support for GP 2016 and GP 2016 R2 ended July 14, 2026. GP 2018 runs out January 11, 2028. Two separate companies told me last year that they had until 2029, and in both cases a ten-minute look at the version number said they were unsupported while we were sitting there talking about it.

The One-Page, Four-Quarter Roadmap

Now you have dates. Put them on a page.

Four columns, one per quarter. Four rows, and resist adding a fifth. Each cell gets a decision, not an activity. “Evaluate data warehouse options” is an activity, which is why it will still be sitting in that same cell a year from now wearing the same confident verb, having cost nothing and settled nothing. “Pick the reporting platform and sign” is a decision with a corpse if you miss it.

That last row is the one people fight me on hardest in the room, and it is also the most valuable row on the page by a distance that is not close. A roadmap that only says yes is a budget request. The not-doing row is what lets a CFO tell a board member no in March, in writing, without having to relitigate the entire thing from scratch in June when it comes back up.

One page. If it does not fit on one page, you are describing the work instead of deciding it.

Which Lines Need a Fractional CIO, and Which Need Contract Staff

This is where the roadmap stops being a document and starts costing money, so be precise about it. Two different problems, two different hires, and people mix them up constantly.

A fractional CIO is for the lines where the decision is the hard part. Vendor selection. Architecture. Telling a board that the integration everyone loves is going to burn ten times its licensed API allowance. I have had that exact conversation, with arithmetic, and the arithmetic ended the debate in about four minutes. A unit-level design we scoped on a regulated manufacturing build would have consumed roughly 1,350,000 API calls a year against a licensed allowance of 130,000, which is over by a factor of ten and not a rounding error anybody could argue their way out of. Batching the same data at the transaction level, around 141 documents a month at four to six calls each, used under a quarter of the allowance. Nobody needed to have an opinion. Somebody needed to multiply.

Contract project staff are for the lines where the decision is already made and the work is real. A data migration. A build. Parallel close support. Hypercare for eight weeks after go-live, then gone.

I am biased toward the fractional CIO model and I will tell you where it is the wrong answer. If your roadmap is four lines and all four are already decided, you do not need a strategist. You need people who can finish things. Hire the project team and skip me.

The Number That Makes This Decision for Most Companies

The Bureau of Labor Statistics puts the median annual wage for computer and information systems managers at $175,140 as of May 2025, with the 90th percentile near $297,510 and employment projected to grow 16% from 2025 to 2035. Loaded, a real CIO in a competitive market is a $300K line item before you count the first project.

At $180M in revenue that math usually does not work. That is not a knock on the role. It is the reason the fractional model exists at all. Our clients mostly cannot afford me full time, and as a slice of a week I still move more than a full-time generalist would, because the slice gets spent on the four decisions that matter instead of on the forty that do not.

For reference on the other side of the ledger, blended consulting delivery runs about $215 an hour in our shop. So roughly 1,400 billable hours buys you the same spend as one loaded CIO salary. You will not use 1,400 hours of strategy. You might use 200, and put the rest into people who build.

Where These Roadmaps Fall Apart

Four failure modes, and I have caused at least two of them personally.

The roadmap has no owner between meetings. A document owned by a committee updates at the speed of the next committee meeting. Put one name on it. Not a department. A name.

Scope creeps in through the back door on the lines nobody is watching. The data migration gets a “while we are in there” request, then another, and by week six the migration has become a chart-of-accounts redesign that nobody approved or budgeted. Our own change-control process exists specifically to stop this, and the rule that does the work is that changes are never bundled to avoid assessment. Bundle three small ones and you have hidden a big one.

Validation and testing get treated as padding. In regulated work, validation runs about 30% of base build hours, and I have watched a buyer pick the competing proposal that quietly left it out because the bottom number looked better on a one-page comparison someone’s analyst made. A materially cheaper proposal is usually cheaper by exclusion, not by efficiency. That is true outside of pharma too. It is just cheaper to learn the lesson there.

The fourth one is the quiet one. Nobody budgets for the year after. Keeping a validated integration healthy across a vendor’s release cycle ran 120 hours and about $31,000 a year on a recent build, and ERP platforms ship on their own schedule whether or not you staffed for it. GP alone takes three all-inclusive updates a year. Put a maintenance line on the roadmap in Q4. Every time.

Gartner put worldwide IT spending at $6.37 trillion for 2026, up 14.2%, and then John-David Lovelock said the part worth keeping. “Despite the strong growth in spending, this is not a rising tide lifts all boats market trend.” Spending is going up everywhere. Whether yours buys anything is a sequencing question, which is the whole point of the page you just built.

Questions I Get in the First Meeting

Nobody here owns IT. What does one afternoon actually buy us?

A usable first draft, because the dates do the prioritizing for you. Collect renewals, published lifecycle dates, and go-lives you already promised a board. That is the whole exercise.

The reason it works is unglamorous. You are not deciding anything yet, so nobody has to defend a position, which means the meeting does not turn into a meeting. Strategy comes after the dates are on paper. I have never once seen it work in the other order, and I have tried.

How far out should this thing go?

Four quarters in detail, eight in pencil. Anything past two years in a mid-market company is fiction, because an acquisition or one genuinely bad quarter rewrites the back half and nobody bothers telling you.

Our CFO already owns IT. Is that actually the problem?

Mostly the wrong worry. CFOs are good at dated commitments and tradeoffs, which is most of what a roadmap is, and I would stop apologizing for the org chart.

The real gap is narrower than people think. It is technical judgment on the three or four calls a year that cost a fortune to reverse, which is maybe twelve days of work spread across twelve months. That is the entire argument for a fractional CIO, and it is also why hiring a full-time one to do it is such an odd way to buy twelve days.

Can our ERP partner just build it for us?

Sure. Then read the byline. A partner’s roadmap is accurate about their own platform and conspicuously quiet about everything competing with it, which is gravity more than dishonesty.

Get the document anyway. It is free, and it is usually well made. Then hand it to somebody with no license to sell you and have them read it sitting next to your actual renewal dates.

If we only do one thing, what is the first line?

Find the auto-renew clause with the shortest notice window and deal with that one. Boring. Usually worth real money. And it proves a roadmap can change an outcome instead of describing one.

Momentum on this stuff comes from one visible win. Not a kickoff deck.

Once the roadmap says go, how fast can we get people?

17 days, on average, for an IT role through KORE1, with a 92% twelve-month retention rate on those placements. Staffing is rarely what holds a roadmap up.

What holds it up is that nobody wrote down what the role was supposed to finish. A req that says “ERP analyst” gets you resumes. A req that says “own the data migration through parallel close, then hand off” gets you somebody who leaves on purpose when the work is done, which is what you actually wanted.

Go Look at Your Contracts

Not the strategy session. The contracts.

Pull every order form, every support agreement, and the lifecycle page for every platform you run. Write the dates in one place. You will find at least one thing you had no idea was expiring, and you will almost certainly find one renewal you have been paying every year out of pure institutional habit. That list, sorted by date, is a better roadmap than ROADMAP_FINAL_v4.docx ever was, and it took an afternoon. The next pass over the same contracts is application rationalization, which asks what each one actually buys you.

Then decide what each date forces, and staff accordingly. The decisions go to someone who has made them before. The build goes to project-based staff you can scale down when it ships. If you want a second opinion on the sequencing before you commit a budget to it, the ERP readiness assessment covers the systems side, or just hit me up through KORE1 and we will go through your dates.

Nice. Now go open the contracts folder.