0330 311 2820
You know the meeting.
Third Tuesday of the month. Someone has printed a pack that runs to fourteen pages. There's a slide with eleven gauges on it, most of them green, two of them amber, and one that has been red for so long that nobody comments on it anymore. Your operations director talks through the numbers. Your sales director looks at his phone. Forty minutes later everyone files out and goes back to doing exactly what they were doing before.
And the thing is, none of those numbers were wrong. That's what makes it so frustrating. The data was fine. The meeting was fine. It's just that nothing actually changed as a result of it.
I've sat in a lot of those rooms. Family-owned engineering firms in the Midlands, food producers in Lincolnshire, a plastics business near Preston that had the most beautiful production dashboard I've ever seen and absolutely no idea why their margins had slipped four points in two years. Different sectors, same underlying issue. Plenty of measurement. Not much movement.
Which is usually the point where someone reads an article about OKRs or hears about them from a mate who runs a software business, and suggests you try them. And then the argument starts. Aren't those just KPIs with a different name? Do we need both? Is this another thing we'll do enthusiastically for one quarter and then quietly stop?
Fair questions. Let's go through them properly, because the difference genuinely matters, and because the honest answer to "which one do we need" is probably not the one you're expecting.
Start with the one you know.
A key performance indicator measures the health of something that runs continuously. It's a vital sign. Your OEE, your on time in full delivery percentage, scrap rate, first pass yield, stock turns, gross margin by product line, days sales outstanding, accident frequency rate, absence percentage. These things existed last year, they'll exist next year, and your job is to keep them within an acceptable range.
That's it. That's the whole concept. A KPI doesn't tell you where you're going. It tells you whether the engine is running properly while you go there.
Good KPIs have a few qualities. They're stable enough to compare across months. They're owned by somebody specific. They're leading rather than lagging where possible, because knowing your rejection rate rose in June is a lot less useful than knowing your incoming goods inspection failures rose in May. And crucially, there aren't very many of them.
That last point is where most firms in the £5m to £50m bracket come unstuck. I once worked with a subcontract machining business turning over about £14m who proudly showed me a monthly pack with sixty three tracked metrics in it. Sixty three. When I asked the managing director which three he'd look at if he could only see three, he went quiet for a long time and then said, honestly, probably none of them, he'd just walk the floor.
He was right, by the way. His instinct was better than his dashboard. Which tells you something about the dashboard.
The second problem is subtler and it's the one I'd really like you to sit with, because it affects almost every firm this size. Your KPIs might be measuring the wrong thing, or measuring the right thing badly.
Take OEE, since it's the number manufacturers love most. The commonly cited figures are that the median plant sits near 60%, the top quartile around 75%, and world class is about 85%, with discrete manufacturing averaging somewhere in the mid-sixties. Only around 6% of manufacturing organisations ever reach that 85% mark. So if your board pack says you're running at 82% and you've never invested in automated capture, you should probably be suspicious rather than pleased.
Here's why. Manual OEE tracking typically overstates the real number by somewhere between eight and fifteen percentage points. Micro stops don't get logged. A three-minute jam while someone clears a chute vanishes. Cycle time assumptions are optimistic because they were set when the machine was new. Nobody records the twenty minutes lost waiting for the forklift. None of this is dishonesty. It's just that a human with a clipboard cannot see what a sensor sees.
I've watched a company install proper capture on one line and see their OEE "drop" from 78% to 61% overnight. The plant manager was furious. He thought the system was broken. It wasn't. He'd been managing a fiction for six years, and the moment he got an honest baseline, he found £180,000 of annual capacity sat in changeover losses he genuinely didn't know he had.
That's the uncomfortable truth about KPIs in this sector. It's not that manufacturers don't measure. It's that a lot of the measuring happens in spreadsheets built by someone who left in 2019, fed by numbers typed in at the end of a shift by an operator who's estimating.
And it's not just you. Only about 7% of UK manufacturers have fully adopted advanced digital technologies. Most SME manufacturers are still running on disconnected legacy systems, losing something like 45 hours a month just to keeping the old stuff working. Government research into SME digital adoption found a clear gap between aspiration and execution, with plenty of firms overrating how ready they actually are.
So before we get anywhere near objectives and key results, ask yourself the awkward question. Do you trust your numbers? Really trust them? Would you bet the value of a new machining centre on your scrap figure being accurate?
If the answer is no, that's your actual project. Everything else is decoration.
Right. Different animal.
An OKR has two parts. The objective is a plain English statement of something you want to be true by the end of a period, usually a quarter. Not a metric. A sentence. Something like "become the supplier our top three customers can't afford to lose" or "get our new powder coating line earning its keep".
Then you attach two or three key results. These are measurable, they have a start point and a finish point, and they describe an outcome rather than an activity. Not "install the new scheduling system". That's a task. It's "reduce average quoted lead time from 26 days to 18 days". If you hit that, something real happened. If you install the system and lead times don't move, you didn't achieve anything, you just spent money.
The distinction sounds pedantic until you see how often it gets missed. Analysis of nearly eight thousand key results written by real teams found that 52% of them were tasks or KPIs wearing a disguise. More than half. People write "hold monthly quality meetings" and think they've set a goal.
Let me give you a worked example, because abstract explanations of this stuff never land.
Say you're a £12m sheet metal fabricator. Margins have been drifting. You've won a lot of low value repeat work that keeps the machines busy but doesn't pay well, and your best press brake operator is retiring in eighteen months with nobody behind him. That's a fairly typical picture for this sector right now.
Your KPIs stay exactly where they are. OEE, OTIF, scrap, gross margin, quote conversion, headcount turnover. Those keep ticking over in the monthly pack.
Your OKR for the quarter might be:
Objective: Stop letting our cheapest work eat our best capacity.
Key result one: Increase average gross margin per shop hour from £38 to £46. Key result two: Reduce the proportion of turnover coming from jobs under £500 from 31% to 20%. Key result three: Get two operators signed off as independently competent on the brake, from zero.
Notice what's happened there. Those aren't things you monitor forever. Once you've moved margin per hour to £46, that becomes the new normal and you go looking for the next constraint. They're change metrics, not health metrics. They have an expiry date. And they force a genuinely difficult conversation, because getting from 31% to 20% on small jobs means telling some customers no, and your sales director is going to hate it.
That's the real function of OKRs, I think. They're not a measurement system pretending to be a goal system. They're a device for forcing a leadership team to agree, out loud, on what matters most for the next ninety days and what they're prepared to give up to get it.
Most management teams never have that conversation. Not because they're lazy, but because it's genuinely unpleasant and there's always something more urgent.
Bridge the Gap between Boardroom Strategy and Shop-floor Reality.
The Implementation Engine brings relentless focus and accountability to manufacturers, turning daily chaos into unstoppable momentum. You could be leading a business that feels different. A team that's united, energised, and crushing goals. Click here!
If you want it in one line: KPIs monitor the business you've already got. OKRs build the business you want next.
I sometimes describe it as the difference between your bank statement and your savings plan. Your bank statement tells you the state of things. It's essential, you'd be reckless to ignore it, and reading it more often doesn't make you richer. Your savings plan is a deliberate choice to move something from where it is to where you want it, and it involves sacrifice.
A few practical distinctions that follow from that.
KPIs run continuously, OKRs run in cycles, usually quarterly. KPIs should mostly be stable, OKRs should mostly change every cycle. KPIs are owned by function, OKRs are usually owned by the leadership team collectively, because the interesting problems in a manufacturing business almost never sit inside one department. KPIs answer "are we alright", OKRs answer "what are we actually doing about it".
And here's the bit that trips people up. A number can be a KPI in one quarter and a key result in another. Scrap rate is a KPI, you watch it forever. But if scrap has crept to 6% and it's genuinely hurting you, then "reduce scrap from 6% to 3.5%" becomes a key result for one quarter, you throw resource at it, and when it's done it goes back to being a KPI you simply monitor.
That's not a contradiction. It's the whole point. The framework isn't about categorising metrics, it's about deciding where your limited attention goes.
Which brings me to the failure data, because I'd rather you heard it from me than found out the hard way.
Depending on which study you read, somewhere between 60% and 70% of organisations that adopt OKRs either abandon them within a year or never get meaningful results from them. That's not a small number. And the most commonly cited reason is exactly the confusion this article is trying to clear up: firms write OKRs that are indistinguishable from the KPI dashboard they already had, then wonder why nothing feels different.
You've just relabelled your monthly pack and added a meeting. Of course it didn't work.
Here's the diagnostic I use. It's crude but it's usually right.
Symptoms. Different people quote different figures for the same thing. Your OEE comes from a spreadsheet. Nobody can tell you the true margin on your top ten customers without a fortnight of digging. Month end takes eleven working days.
You need KPIs. Properly. Not more of them, better ones. Pick six to eight that genuinely matter, get honest capture on them even if it means spending money on sensors or a decent ERP module, and accept that the first accurate baseline is going to look worse than the fiction you've been living with.
Do not start OKRs. You'd be setting targets against numbers you can't measure, which is just organised guessing. If you want help with the cost of this, Made Smarter is worth a look. It's invested over £112 million in direct grants and the SMEs who've gone through it report average productivity improvements of around 26%, which is not nothing.
Most firms between £5m and £50m are in scenario one. I'd guess two thirds. I know that's not the exciting answer.
Symptoms. The dashboard is accurate. Everyone knows the problems. The same three issues have been on the risk register for two years. Your improvement projects get started and then get deprioritised the moment a big order lands.
You need OKRs. This is precisely what they're for. You've got the diagnosis and no treatment plan. Three objectives, maximum, for the quarter. Kill everything else.
Symptoms. Fourteen live improvement initiatives. A new ERP, a Made Smarter project, an ISO 14001 push, a rebrand, two new product lines, and someone's cousin doing something with AI. Everyone's busy. Nothing lands.
You need OKRs, but used as a pruning tool rather than a goal setting one. Write down everything currently in flight, then ask which three would you regret not finishing. Stop or park the rest. That's a brutal meeting and it's the most valuable one you'll have all year.
In practice, of course, you end up needing both. KPIs are the floor, OKRs are the ladder. But sequence matters, and I've watched more than one business damage its own credibility by rolling out a shiny goal framework on top of data nobody believed in. The workforce isn't daft. They can tell when senior management is measuring a fantasy.
You don't have spare capacity. I know that. Skills shortages are the number one barrier to growth for UK manufacturers right now, with half saying it's their single biggest constraint and talent gaps being the main thing stopping SMEs growing into larger firms. Nobody in your business is sitting around waiting for a new framework to administer.
So here's the stripped-down version.
Three objectives, no more. For a business your size, honestly, two is often better. Every objective you add halves the attention the others get.
Two or three key results each. The data on this is quite striking. High performing organisations average about 2.9 key results per objective. Struggling ones average 3.5. That gap looks trivial and it isn't, because attention is the scarce resource, not ambition.
One page. Printed. On the wall. If your OKRs live in a system that people have to log into, they don't exist. I'm slightly joking. Only slightly.
Twenty minutes, every Monday. Not a review, not a presentation. Just: what moved, what's stuck, who needs help. The single biggest predictor of whether an OKR programme survives is whether it has a heartbeat. Programmes run by one person alone are dead on arrival about 70% of the time. Programmes owned by a proper team stay alive around 90% of the time. It's not the framework. It's the fact that people show up.
Keep it away from bonuses. As soon as an OKR affects someone's pay, they'll set targets they know they can hit. You'll get 100% achievement and zero progress. Your KPIs can drive reward if you like. Your OKRs shouldn't.
Expect the first cycle to be mediocre. Teams in their first two cycles average around 51% completion. By cycle five, the same organisations average 79%. Most firms give up somewhere around cycle two, which is precisely the point at which the data says it's about to start working. Two or three cycles is the realistic window before you see something meaningful, so you're looking at six to nine months, not six to nine weeks.
That last point is the one I'd underline. If you're not prepared to give this a year, don't start. You'll just add another abandoned initiative to the pile, and the next time you propose something the shop floor will remember.
So, back to the question in the title.
KPIs tell you how the business is running. OKRs decide what you're going to change about it. They're not competing systems and choosing between them is a bit like choosing between your fuel gauge and your destination. You need both, but only one of them is going to get you anywhere.
If your numbers are shaky, fix the numbers first. That's dull, unglamorous work and it's almost always the right answer for a firm in this bracket. Six to eight metrics you'd stake money on beats sixty three you half believe.
If your numbers are good and you're still standing still, you don't have a data problem, you have a focus problem. That's what OKRs are for, and the discipline is less about the format than about being willing to say no to eleven things so that three can actually happen.
And if you take nothing else, take this. It's the test I keep coming back to after years of watching these things succeed and fail. Whatever system you use, it only counts if it changes what somebody does on a Tuesday morning. If your framework doesn't alter behaviour on the floor, in the office, in the way a supervisor decides what to run next, it isn't a management system at all. It's just paperwork with a nicer font.
Manufacturing output in the UK is growing slowly, productivity is going backwards once you strip out inflation, and the firms that come out of the next few years in decent shape will mostly be the ones that got deliberate about a small number of things rather than busy about a large number of them.
You've probably already got the dashboard. The question is whether anyone's driving.
Here's the pattern I see again and again in firms your size. It isn't that you lack data, and it isn't that you lack ideas. You've got a folder full of both. What's missing is the machinery that turns a decision made in a Tuesday meeting into something that's actually different on the floor by Friday.
That's the gap our Implementation Engine Programme was built to close.
It's designed for established manufacturers who are past the startup scramble and into the harder problem of making change stick across a real business with real customers and real constraints. We help you get to numbers you'd genuinely bet on, agree the small handful of things that matter this quarter, and build the weekly rhythm that stops it all quietly sliding once the next big order lands.
No fourteen page packs. No frameworks for the sake of frameworks. Just the discipline to finish what you start.
Take a look at the Implementation Engine Programme here
If you're not sure whether you're in scenario one, two or three from this article, that's a perfectly good place to start the conversation. Most people aren't sure. Working it out is usually the first bit of value.
Pay It Forward! Sharing Is Caring!
