Submitted on: September 23 2026

By Fredric, Principal of Bright Plum

You approve a number every month. Can you say what it bought?

Most directors I talk to can't. It isn't because they don't pay attention. It's because nothing arrives that would tell them. The work happens somewhere out of sight. The only signal comes at the end of the year, when somebody mentions the hours are nearly gone.

So here's a month, week by week. You can hold it against the one you're getting now.

Week one: table setting

This is the week that decides whether the other three are any good.

We look at the budget for the month. We look at what carried over: work left unfinished, work still in progress, anything that got deferred and shouldn't get deferred again. We figure out what needs to get done before the next check-in with you. Then the team meets internally and plans the month.

Weeks two and three: the work

These are the heavy weeks. They move depending on when your meeting falls.

The work splits two ways. There's proactive work, which runs against the six dimensions we track on every site: security, performance, accessibility, documentation, communication, and platform health. Some of that needs your approval before we start. We bring it to you rather than doing it quietly and billing for it.

Two things worth separating because vendors blur them. The audit is what we run on a schedule. The remediation is what it costs to act on what the audit found.

Security runs monthly. The scan is quick and the fix is usually bumping packages to versions that aren't vulnerable.

Accessibility runs quarterly because the audit is cheap. The fix, however, isn't. One scan can surface months of work. None of it is a version bump. That's the line item typically left out of a budget.

For almost all of our clients, we've inherited their site. We go slower at the start. We don't know yet how it's meant to work, and often nobody still at the institution does either. So we ask questions and build automated tests. Changing something you didn't know was important is worse than leaving it alone another month.

Then there's what you asked for: a feature, a bug, support for your web team. When those come in, they get prioritized in the same week, and every task gets an estimate and a due date.

The client meeting lands in here. We bring the site report.

Week four: looking back, and deciding together

Finish what's open, then reflect on it with you.

Sometimes work needs doing and the budget won't reach. We borrow ahead against next month, and we get your approval in writing before we do. You should never find out after the fact that a month ran long.

Sometimes you tell us it isn't that important and it can wait. Good. It moves into next month's planning and we refocus on something else.

Most months it's simpler than either. The work is going to land, the budget is going to hold, and we'll have a fuller update for you next month.

The collaboration is the part most vendors skip. They plan, they ask you for a priority, they execute. Then the month ends and nobody comes back to ask how it went.

That's the week that makes the next one work. It also produces something that compounds: a record, month by month, of what the site needed and what got decided. A year of those is a budget case you didn't have to write.

What arrives without you asking for it

The site report.

It's built to be forwarded. Your team can read it and so can your board. It shows what we did, where the site stands on page speed and what changed in your analytics, and what we recommend next.

You approve a number every month. Can you say what it bought?

Past, present, and what's coming. That third part is the one that matters, because the reason you need this isn't curiosity about last month. It's that budget renewal is coming, and you should be making that decision with information rather than a feeling.

The month when nothing breaks

This is the month people struggle to justify, and it's the one you're actually paying for.

"Nothing broke" looks backward. The report looks forward, and in a quiet month it carries more, not less.

Nothing broke, but a security advisory is coming. That goes in the report.

Nothing broke, but we ran an accessibility audit and found pages that need remediation. That goes in the report.

Nothing broke, but the analytics show a group of visitors having a slow experience. We open a ticket and ask for approval to work on page speed.

Nothing broke, and you're current on the things that come first. Then we recommend looking at how people use the site.

That one isn't standard. It's a recommendation, and it competes for the same budget as everything else. If you're behind on versions or a compliance deadline is closing, that work goes first and this waits. But when the question is "nothing broke, what now," this is usually the answer.

Performance isn't only load time. It's whether a visitor finds what they came for, whether the language and spacing hold together, whether a page needs less on it rather than more.

We look at your authors the same way. What are they trying to get done, and how many steps does it take? Sometimes the finding has nothing to do with visitors. Your site works fine for the public and your team is making the same change in four places to publish one thing. Nobody designed that. It accumulated, and it comes straight out of the time your team doesn't have.

What you get from most vendors instead

A report of what was done, if you get one.

Nothing arrives until late in the year, and then it's one of two messages. You haven't used your hours, here's what's left. Or the thing you're asking for is outside your hours, here's what it would take.

Both arrive too late to do anything with.

We forecast instead. Against the budget you already approved, not against a new proposal. And we listen in the monthly check-in for what your team needs. Where are you spending the most time? What would give you time back? The answer to that usually points at the work worth doing next.

What a recommendation should sound like

We don't recommend perfection. We recommend a balance between budget, quality, experience, speed, and priority.

Say performance matters to you, and a few pages are heavy. Those pages don't get much traffic and they hold a lot of content. We could spend real budget rebuilding them properly. Instead we cache them harder, which costs a fraction, and put the rest of the budget into the authoring experience your team touches every day.

The second one is worth more to you. It just doesn't look as impressive on a list of completed work.

Sometimes you'll tell us your priority goes first and ours waits. That's yours to decide. Our job is to make sure you're deciding it with the risk in front of you, in writing, rather than finding out later. The budget is the hard line, not our list of recommendations.

If your vendor's recommendations always cost more than your budget, they aren't recommendations. They're a sales pipeline.

What to ask your vendor this week

Tell me what's coming.

Not what you did. What's ahead in the next six to twelve months, and whether the budget I've already approved covers it.

Then ask the one question that separates a proactive partner from a responsive one: how will you use the rest of my budget to give my team time back?

Your list of sites is going to keep growing. Your headcount isn't. A partner who only answers requests is adding to your workload, one ticket at a time. A partner worth keeping is subtracting from it.

If they can't answer that inside the money you've already spent, you have your answer.