A few years ago one of my team suggested we change the order we packed trade orders for morning pickup. Small idea. Took ten minutes to explain over a coffee. I said yes, we tried it, and everyone agreed it “felt faster.” That was the end of it. No numbers, no follow up, just a vague sense that things were better.

The problem is that a vague sense is worthless when you want to build a business where people keep bringing ideas forward. If I cannot tell you whether that packing change saved us fifteen minutes a morning or nothing at all, I cannot reward it properly, I cannot repeat it elsewhere, and I certainly cannot use it to convince the next person that speaking up is worth their time. Most of the advice online about employee ideas stops at “recognise the contribution” and “make people feel valued.” That matters, but it is only half the job. Recognition without measurement is a pat on the back that both of you secretly suspect might be undeserved.

This article is the other half: a practical method for working out whether a specific idea produced a real, attributable result. You do not need an analyst or a business intelligence tool to run it. You need a metric, a baseline, a checkpoint, and the honesty to report what actually happened.

Why “feeling better” is not a measurement

There is genuine research showing that people care deeply about doing work that has an impact. The American Psychological Association (n.d.) frames making a difference at work as a real driver of wellbeing, and DecisionWise (n.d.) treats seeing positive outcomes from your own work as a core part of engagement. I do not dispute any of that. The gap is that this material treats impact as a feeling the employee has, rather than a number the business can verify.

Both things can be true at once. Your staff member can feel proud that their idea mattered, and you can also have a defensible figure that says it saved four hours a week. The feeling is fuel. The figure is proof. When you only have the feeling, three bad things happen. You start crediting ideas that did nothing. You miss ideas that quietly worked because nobody looked. And over time your team learns that “impact” around here means whoever pitched their idea with the most enthusiasm, not whoever actually moved the needle.

The fix is to decide, before you implement anything, how you will know if it worked.

Step one: define the success metric before you implement

This is the single most common miss, and it is the one that undoes everything downstream. If you wait until after the idea is running to decide what “success” looks like, you will unconsciously pick whatever metric happens to have improved. That is not measurement. That is finding a target around the arrow after it has landed.

So before the idea goes live, write down the one number it is meant to move. Be specific. “Improve customer service” is not a metric. “Reduce the number of complaints logged about late deliveries per week” is. Good candidates in a small business are almost always concrete and countable:

Pick the one metric the idea most directly targets. If someone suggests a new packing order to speed up morning pickup, the metric is pickup time, not overall customer satisfaction. Keeping it narrow is what lets you attribute the result later.

Step two: capture a genuine baseline

Once you know the metric, you need a before figure. This is where a lot of owners fall down, because they realise they have never actually measured the thing the idea is about to change. That is fine. Spend a week measuring it before you change anything. A baseline built from a single lucky or unlucky day is useless, so take a representative window: a fortnight of morning pickup times, four weeks of complaint counts, a month of wastage figures.

You do not need software for this. A spreadsheet, or even a clipboard by the packing bench, is enough at this scale. The Nimble (n.d.) comparison of spreadsheets and dedicated tools is a fair reminder that spreadsheets do eventually break down, but for measuring one metric across a few weeks, a spreadsheet is honestly the right tool. The point is not the tool. The point is that you have a number written down and dated, so that in six weeks nobody can argue about what “before” looked like.

Step three: set a fixed review checkpoint

“Let’s see how it goes” is where good measurement goes to die. Open-ended reviews never happen, because there is never a moment that obviously demands one. So when you implement the idea, put a date in the calendar. Thirty days is a sensible default for most operational changes; sixty if the effect is seasonal or slow to show. Treat that checkpoint as a real appointment, not a maybe.

Setting the checkpoint in advance also protects you from two opposite mistakes. It stops you declaring victory in week one off the back of a good Monday, and it stops the idea quietly running forever without anyone ever asking whether it earned its place. If you are rolling the idea out gradually, the same logic that governs a good pilot applies here. Stackby (2026) and Atlassian (n.d.) both make the point that a defined review window with agreed success criteria is what separates a real test from just doing something new and hoping.

Step four: compare before and after on the specific metric

At the checkpoint, put the two numbers side by side. Baseline pickup time versus current pickup time. Complaints per week before versus after. That is the comparison that matters, and it is far more honest than asking the team whether things “feel” better. People are terrible at estimating this. I have watched a change that everyone swore had transformed the morning turn out to have saved about ninety seconds, and I have watched a change nobody rated turn out to have quietly cut our reorder errors in half.

Resist the urge to swap in a rosier metric if the target one did not move. If the idea was meant to reduce complaints and complaints did not fall, the fact that morale went up is a separate finding, not a rescue. Keep the scorecard honest.

Step five: separate correlation from attribution

Here is the discipline that most feel-good content skips entirely. Your business is not a laboratory. Lots of things change during any measurement window, and the idea only deserves credit for what it actually moved.

So when you sit down at the checkpoint, list everything else that changed during the window. Did you hire someone? Was it a quiet month seasonally? Did a competitor close down the road? Did you change three things at once? If the packing order changed in the same fortnight you also took on a new casual, you cannot cleanly credit the packing order for a faster morning. This is why changing one thing at a time is worth the patience. When you must change several things together, at least name the confounders out loud so you are not fooling yourself. Giving an idea credit it did not earn is just as corrosive as ignoring one that worked, because it teaches you the wrong lesson about what to do next.

Step six: close the loop with the actual numbers

When you have a defensible result, take it back to the person whose idea it was, and take the numbers, not just a thank you. “Your packing change cut our average pickup time from eleven minutes to seven, across the last month” lands in a completely different way than “great idea, thanks mate.” One is a compliment. The other is evidence that this person made the business measurably better, and that you noticed precisely how much.

This is the part that compounds. Tick HR (n.d.) is right that recognising contributions matters for morale, but recognition backed by a real figure does something recognition alone cannot: it teaches the whole team that ideas here get measured and rewarded on merit. That is the culture you actually want. I have written more about turning that evidence into a genuine feedback habit in tracking employee feedback against business results, and about what happens when you skip it in the piece on the employee idea graveyard.

Step seven: report the losses honestly

Not every idea works. Some make no difference and a few make things worse. What you do with those matters more than what you do with the wins. If you only ever celebrate the ideas that paid off and quietly bury the ones that did not, your team learns a precise and damaging lesson: only pitch safe ideas that are certain to succeed. That kills exactly the ambitious, slightly risky thinking you most want from the people closest to the work.

So when the checkpoint says an idea did nothing, say so plainly and without blame. “We tried it, it did not move the number, here is what we learned, thanks for putting it up.” Treat the honest null result as a normal, respectable outcome. That is what makes people comfortable bringing you the next one. An idea that fails cleanly and is measured honestly is worth far more than an idea that “worked” on vibes, because at least you learned something real.

Making this a habit, not a one-off

The reason this is worth building into how you run the place is that it compounds. The first time you run the loop it feels like extra admin. By the fifth time you have a track record: a small library of ideas, each with a baseline, a checkpoint, and a real before-and-after figure. You start to see which kinds of ideas from which parts of the team tend to pay off. You get faster at spotting the metric an idea should target. And your people start framing their own suggestions in terms of the number they think it will move, because they have learned that is how ideas get taken seriously here.

Tracking whether an idea actually shifted a metric, with a defined checkpoint instead of a note that gets forgotten, is exactly the kind of before-and-after visibility I built businessreview360.au to support. It turns a submitted idea into a tracked item with an owner, a target metric, and a review date, so closing the loop with staff becomes the default rather than something you mean to get around to. You can run the whole method on a spreadsheet, and for a single idea you probably should. The value of a system shows up once you are running the loop often enough that remembering every open checkpoint by hand becomes the thing that quietly breaks.

References

American Psychological Association. (n.d.). Employees really value making a difference at work. Here are 7 tips to help them. https://www.apa.org/topics/healthy-workplaces/making-difference-at-work

Atlassian. (n.d.). How to create a rollout plan for a project launch. https://www.atlassian.com/agile/project-management/rollout-plan

DecisionWise. (n.d.). Impact: Seeing positive and worthwhile outcomes and results for your work. https://decisionwise.com/resources/articles/impact-seeing-positive-and-worthwhile-outcomes-and-results-for-your-work/

Nimble. (n.d.). CRM vs Excel: Why small businesses outgrow spreadsheets. https://www.nimble.com/blog/crm-vs-spreadsheets/

Stackby. (2026). How to run a successful pilot project (strategy + steps). https://stackby.com/blog/pilot-project/

Tick HR. (n.d.). Recognising your employees’ contributions and ideas. https://www.tickhr.com/employees-contributions-and-ideas/

FAQ

How long should I wait before measuring whether an employee’s idea worked?

Set the checkpoint before you implement, and for most operational changes thirty days is a sensible default. Use sixty days if the effect is seasonal or slow to show, such as anything tied to a monthly billing cycle or a quiet-versus-busy trading period. The key is that the date is fixed in advance and treated as a real appointment, not an open-ended “let’s see how it goes” that never actually arrives.

What if I never measured the thing before the idea changed it?

That is normal and easily fixed. Spend a week or two measuring the metric before you change anything, so you have a genuine baseline. Take a representative window rather than a single day, because one lucky or unlucky day tells you nothing. A dated figure in a spreadsheet is enough. Without a before number you have nothing honest to compare the after number against.

How do I know the idea caused the improvement and not something else?

You separate correlation from attribution. At the checkpoint, list everything else that changed during the measurement window: new hires, seasonal swings, other changes you made at the same time. If you changed three things at once, you cannot cleanly credit any single one. This is why changing one thing at a time is worth the patience, and why naming the other factors out loud keeps you from fooling yourself.

Should I tell staff when their idea did not work?

Yes, plainly and without blame. If you only celebrate the wins and quietly bury the misses, your team learns to pitch only safe ideas that are certain to succeed, which kills the ambitious thinking you most want. Treat an honest null result as a normal, respectable outcome. An idea that fails cleanly and is measured honestly is worth more than one that “worked” on vibes alone.

Do I need software to run this, or is a spreadsheet enough?

For a single idea, a spreadsheet is genuinely the right tool, and you should not overcomplicate it. The method matters far more than the tool. A system starts to earn its place once you are running the loop often enough that remembering every open checkpoint and baseline by hand becomes the thing that breaks, at which point tracking each idea as an item with an owner, a target metric, and a review date saves you from ideas quietly slipping through the cracks.