Why some of your projects take ten times too long (and what to do about it)

Last updated: February 2019

In the following talk, we describe how to make your company move at least ten times faster.

That may sound like an overstatement, but bear in mind that we’ve seen one company take six months to complete a task that took another company thirty minutes.

That’s 8,760 times slower.

We love the techniques that we describe in the following talk. They work amazingly well. One of our current clients, when they first joined us, was taking 119 days to turn an idea into a statistically significant win. We reduced that time to 7 days. And increased their sales by $29 million per year.

We hope you find the techniques as useful as we do.

The slides

The slides from the talk, some of which may be confusing without the audio (see the video below).

A video of the talk


A video of the talk.

The talk in podcast format

To download the audio of this talk, and others, subscribe to our podcast.

An unedited transcript of the talk

What I’m talking about now is why some of your projects take ten times longer than they should, and what to do about it. Now, let’s just take a look at this little car on the right there, that little red car. It can do over a hundred miles an hour. The driver has been driving for more than seven years. Very qualified driver, very adept driver. That little car over there, again, can easily do over a hundred miles an hour. The driver’s been driving for 15 years. This one here, the driver used to be in the police. Amazing driver, did an advanced driving course and is driving a car that can do over 130. This is a bus driver. Drives for a living.

All these people are amazing drivers, they’ve got good cars, so why is it that none of them are actually moving at all? Why are they not moving? They’re in a traffic jam, and no matter how skilled they are, no matter what tools they’ve got, none of them are moving because the roads that they’re moving on, the process they’re following has terrible bottle necks, hasn’t been well enough designed. And so everyone’s skilled, no-one’s actually getting anywhere.

And that’s what my talk’s about. My talk’s about how so many companies have got a lot of very bright, keen, motivated people who are working until eight o’clock every evening, they’re all really ambitious, they’re very smart. They know a lot about web marketing and web design and conversion and usability. They’re using all the right tools, they’re tracking, managing their tasks using getting things done methodology. They keep storing their tasks in OmniFocus and Basecamp and lots of other things. So why is it that projects move so slowly?

The fact that I find incredible is that we know companies that have taken 30 minutes to implement split testing software. We also know companies that have taken six months to implement split testing software. Exactly the same job, but one of them is 8,760 times faster than the other. It doesn’t take a rocket scientist to work out that that company on the left has a massive advantage over the one on the right. Often the ones on the right are larger companies, but not always. Not always. So this is what the talk is about today. How do you know if this problem pertains to your company is if you feel that everyone’s productive but everything takes ages to happen, if you feel as if things take longer to happen than they seem to be, if you think: hang on, implementing that split testing software, it’s only two lines of code, can it really be that difficult to paste it in some documents that we actually own? The web pages? And also that some projects never finish. If you find that within your company there are lots of projects that were started six months ago or a year ago and which still, for some reason, aren’t actually finished yet, aren’t published yet.

Here’s just a little background about our company because I think it explains how we know about this. This first picture is an aerial view of a processing plant. I used to be a scientist. I did a PhD in high temperature materials for lining combustion chambers. And then I moved on to become a process engineer where I learned a lot about optimizing large-scale industrial processes. This plant was over 100 meters long, and there’s a massive amount of money involved when you’re optimizing processes that are running 24 hours a day with so much resting on it. They are lorries, I think, those things lined up there just to show the scale of it. And so normally, when we talk about web marketing, we often talk about it in terms of process science, process engineering. Today, the process we’re talking about isn’t the process of turning your visitors into customers. Today the process we’re talking about is the process of turning your ideas into actual money, into finished web pages which are generating money for you.

My co-founder and I originally used to work in-house like many of you, and we tripled the sales of the business using conversion rate optimization within 12 months. Over a period of three years we grew the company’s sales by 12 times. And one thing that we weren’t so aware of then is how agile, how fast-moving that company was. Because lots of our processes were very quick, we could implement things amazingly quickly. And that’s what I want to help you out with and give you some very useful techniques and tools that you can take straight back and start applying to your own business. What I’m hoping is that you’ll look back to your own business at the end of this talk, and you suddenly start spotting things that have been staring you in the face for a long time that you’ve been doing wrong.

We wrote an article about what we did and Google took notice. They invited us to become the first Google partners for Google’s testing tool. Now, we’ve worked with clients with websites in 9 languages in 22 countries. So we’ve had an amazing amount of insight into what the differences are between the highly functional, highly fast-moving companies and the others. And we’ve worked with some really fast-moving companies: Apple, Google, Facebook, Amazon, and lots of smaller businesses. And we’re always focusing on getting results, so it has been a very good way to learn about what works in web marketing. And what you’ll find is Amazon’s not just a company that’s gone a long distance and has got huge economies of scale, but they are also implementing extremely quickly even now. The real strength of many successful web companies is that not only are they blazing ahead but—back to that road analogy—they’re traveling much quicker. They’re in the fast line while lots of other companies are still in traffic jams.

So how should you create processes? How can you create processes that deliver value efficiently? The first point: today a manager is analogous to the designer of a road network. And it used to be that a manager was the person with a task list who used to walk around telling everyone what to do, bossing people around, keeping things in order. These days software can do a lot of that. Now, a manager is someone who can design workflows in a way that makes it easier to carry out, more foolproof. As a manager, you need to be thinking about learning how to design efficient workflows. And that involves knowing about queuing theory, and knowing about the process engineering techniques that I’m going to cover now.

You must be able to build processes and that means you need to understand or identify what the customer values and then you need to build a process in which each step is a supplier to the next step and a customer of the previous one. You need to think of all of your workflows, even within your own company, as processes. The person who finally integrates the page or who codes up the page is a customer to the graphic designer. And the graphic designers is a customer is the person who was before them in the chain. So steps pull work. So many companies still push work through the system. Someone will come up with an idea, push it to the copywriter, push it on and you end up with traffic jams in that case. So you should be thinking in terms of pull.

Here’s a process that will make that easier to understand. At the top, there are two elements: the green boxes and the yellow boxes. The green box is what you call a resource, it’s a value giver. That’s in a very abstract way. In terms of engineering, it could be a machine or a person or information. If it were a machine, it might be a car-assembling robot and this orange thing could be a piece of steel that’s gradually being processed and turned into a car. But for all of the rest of this talk, I’m going to assume that the resource is a worker, a person—for example, a conversation practitioner. And the flow unit is not steel, it’s information. The information starts off as an idea, and as we work from left to right, the idea gradually gets turned into a webpage and finally put onto the website.

Now ideally with the process, neither the worker nor the work unit would be sitting around. Because you don’t want either of them doing nothing. Ideally, you’d have it so that this person second from the left, who, let’s say, would be the copywriter, you’d like it so that the moment the copywriter has finished working on this bit of work, that they pass it on to the next person, and at that same moment, the next idea goes straight to the copywriter. So the work flows constantly and there are no pile-ups. The problem with that is if there’s any variation in the flow of work units. Unfortunately, there always will be variation, because everyone takes different amounts of time, and some bits of work will be problematic. So one of them must sit around—either the worker needs to sit around or the piece of work needs to sit around. Which should it be?

I think you’re going to be surprised at the next four slides.

I’m going to tell you about timelines. Worker efficiency is the percentage of time that a worker spends delivering value—that is, working. There’s the little worker. If we were to say to the worker, “How did you spend your last few weeks?” then the worker might produce this chart. Green represents times that they were working on something. Red represents time that they were waiting—just doing nothing, waiting for work to come in. In most companies, you’ll find that even this is an understatement. And so the worker efficiency is the time spent giving value divided by the total time. You’ll probably find that it’s extremely high for your company, because most are. The worker spends almost all of their time adding value.

Work unit efficiency—as opposed to worker efficiency—is something that most people haven’t even thought of. It’s the percentage of time that the work units spend receiving value. So there’s the little work unit, the little wireframe sitting there. And if it could talk, if we could ask it, “How did you spend your last few weeks?” it might have a different story to tell. It might say, “Well, someone did a little bit of writing and I turned from an idea to a first draft at this point and then I sat around for a long time and then on Thursday afternoon someone worked on me for two hours and then again I sat around and I was on Leslie’s desk, in an in-tray, and then finally she did this little bit.” Most of the time, the work units are sitting around doing nothing. So if a web design, this is the “webpage’s-eye view.” And the work-unit efficiency is the time spent receiving value (that is, being worked on), divided by the total time from start to finish (from the start date and time to the finish date and time—the calendar time.

If you maximize the worker efficiency, then the workers are treated like royalty and the work units have to wait. Now, this looks like a typical processing plan, doesn’t it? Where you can see that this person here, this worker, is working on this one but there are two waiting there. And then the second that they finish this bit of work, it’ll go into a queue waiting for the next person. So that’s what you’d call maximizing worker efficiency. If you imagine going through an airport, you know what it’s like. This is the passport control person, this is the security person, in the work units are the travelers who are going to fly. And the travelers spend most of their time waiting for the passport check people and security people who are treated like royalty, and who keep them waiting. The same happens if you go for an operation. You’ll probably find that you go to the doctor then you see a specialist and then you have an operation. But there’s a lot of waiting in between, because those people are maximized in terms of their efficiency and in terms of their busyness.

To keep the workers busy, each worker must have an uninterrupted supply of work units. That is the next work unit needs to be ready as soon as the worker becomes ready. So you need a buffer, a queue of work units waiting to be processed. But that lowers the work unit efficiency, obviously.

Another approach would be to maximize the work-unit efficiency, in which case the work units are treated like royalty and the workers have to wait. Which looks really weird, doesn’t it? It’s strange the idea that this wireframe designer is now just sitting around waiting, doing nothing, while waiting for the work to come in. An example of work-unit efficiency would be medical care for a President. The President goes to the hospital and you know what happens. The medical team members are waiting. The President doesn’t arrive at the hospital and then wait around for six hours like many of us would, and then finally gets seen by a doctor, and then is left in a corridor for an hour and then the operation happens. With a President, the health workers are all ready and waiting. The same happens when a head of state visits a foreign country. If the Queen goes to visit a country, and she’s being taken on a tour of a car plant, she’s not going to have to wait half an hour for the car plant workers to be ready. They’ll just be standing there waiting for her to arrive.

In web terms, you might think of work-unit efficiency as prioritizing a project. It probably looks quite unfamiliar to you. So to summarize, to get high work-unit efficiency, each worker must be ready as soon as the work unit becomes ready. So you need the workers to be waiting. But that, by definition, lowers the workers’ efficiency.

So which type of business should yours be? Should it be high worker efficiency or should it be high work unit efficiency? Most people would assume that the answer is high worker efficiency. Workers are the people we are short of, they cost us money, and we need to make the most of them, so let’s maximize their utilization (that is, keep them as busy as possible). It’s not a problem if a wireframe sits around for a few days. It is a problem if a graphic designer sits with his or her feet up on the desk.

But here’s the paradox of worker efficiency. Worker efficiency creates a lot of problems in a company—what we call secondary needs. The problems arise from the fact that the work units are sitting around. You may wonder why that’s a problem. Is it really a problem? A wireframe is hardly going to complain or cause trouble while it’s sitting around. But here are some of the many insidious ways in which that causes problems.

First, the work units need somewhere to be stored, they need to be stored on task lists or in people’s heads. So what you need then is you need a system for storing large numbers of tasks, so you need, like we mentioned before, some kind of task management system. But that’s just the start of it. Then we need a way of retrieving them, so once something’s gone into the task list, we need a way of pulling it back out. You need to do re-work. For example you need to re-read material to bring yourself back up to speed. Have you ever been in a meeting where pretty much everyone in the meeting can’t even remember what the meeting was about? At the start everyone’s frantically reading the notes from last time because no-one can remember what we’re even meeting to do. And everyone’s preparing for the meeting because it takes 15 minutes to remember where we got up to last time, a month ago, when the project was happening.

Amid all the mess it’s hard to see the obvious, you lose the clear view of the business. It’s possible, we’ve known companies where there’s so many tasks, there’s so many task lists all over the place that projects have literally gone missing and someone will just say, “Whatever happened to that test we were going to run on the Tell a Friend program?” And someone else will say, “Oh I don’t know, we’ll have to look into that.” And they’ll rummage round and find it. So it’s amazing how, quite quickly, you can get to that you can’t really see what’s important in the business, what’s going wrong? There’s so much … it’s like looking at a congested road network where it’s actually hard to see where the bottlenecks are.

You get work that only exists because we chose to work this way. Like planning for meetings, for example. The window of opportunity closes, for example a key person leaves the company, the project had been running for over a year. A new check out process or whatever. And the project ends up … no-one even knows how it’s happened or what’s happening. Your customer, for example your boss, colleagues or client become impatient and unhappy and want reassurance. So just actually providing that reassurance takes time, that secondary work. The work unit itself develops secondary needs. The work unit … maybe it goes stale. An article that was nearly ready to be published is no longer relevant or it needs re-working. The missed opportunity. Sometimes by the time a project’s finished, we’ve seen projects where, by the time they’re ready, the moment has passed and it’s not really something that can be done anymore.

Here even more, problems arising from the fact that the work units are sitting around continued. Work in progress is money that’s tied up. This is one thing that’s obvious if you work in a car processing plant and you’ve got, I don’t know, 300 tons of aluminum all sitting there in the warehouse and that’s all been processed and bought but it’s less obvious when you’re working in information like we all are. So the most harmless type of to-do is a some day maybe. It’s one that’s an idea that you stick on a to-do list and say, “Some day we could do that but for now let’s not work on it.” The most harmful type of to-do is one that’s already had money spend on it. So imagine that each work unit, everything on your task list that’s already started, if you look at all your tasks and think, what’s the dollar value? How many man hours or person hours, sorry, of work have been done on that so far, that investment only pays out when the work unit’s published, so when it reaches the end of your process. So if you’re actually to put a dollar value on all of the work in your to-do, it can be quite upsetting because you think, wow, we’ve had … people have worked weeks on this project and it’s still not shipped and we’re actually … currently someone is starting another project before this one’s even finished.

Also challenge the checked off task. Once a task is complete, often a huge hidden type of work in progress is work not fully exploited. We know someone who created a video for a charity, spent about a month traveling to the charity and putting together all the footage, finally got it approved and then, I think it was about two years later, realized that the charity had forgotten to actually make any use of it and only 500 people on YouTube had seen it. And so, yes, in a way that was published, but it wasn’t actually … it had reached the end of one process but hadn’t gone all the way into being fully exploited. So yeah, 500 people viewing a video that was a months worth of work.

The golden rule with lots of this is keep your WIP, your work in progress, low. You can solve a lot of the problems within your company that I’ve talked about by keeping the work in progress low. Minimizing the amount of new projects that get started before old projects get finished. There is one argument to say that … I spoke to a business owner last week and said, “Do you ever start projects before an existing project has finished?” And he said, “Oh yes, we do that all the time.” And I said, “Why? Why? Did you just get bored? Did you think you weren’t busy enough?” And so often it’s a real problem in companies because it’s so easy to start a project without remembering how much value in terms of work in progress is already in the chain.

A good question for how to detect secondary work is consider, or even ask if you dare, whether your customer would willingly write a cheque to pay for this activity. So whatever you’re working on, if you think, am I actually working on work or am I working on secondary work that’s created by the work? And if you ask, would the customer pay for this? Often you’ll think the customer would be horrified to know I was working on this, nevermind pay for it. So that’s a good way to spot when you’re doing that work that’s only created because of your process.

What causes low work unit efficiency? So work sitting around. To-do lists and task management software lure us into storing work. So because … in a processing plant when the plant manager comes around and sees a massive warehouse full of half-made parts, they’re kind of understandably angry because they can see that that’s a waste of money and they can see all the reasons why it’s a waste. I wouldn’t say that web marketing managers or web design managers have that same level of understanding when they see a large task list. They don’t have that same understanding that that’s just a bad thing. A large task list is a bad thing. And also, to-do lists are like warehouses. They hide problems. It’s often very hard to spot where the bottle necks are if those buffers exist. In companies like Toyota, they don’t have warehouses in between because if part of the process isn’t keeping up with the rest, they want to be able to see it. And so the whole line comes to a stand-still if one part of the line breaks down. That visibility they have is incredible. Another thing to keep an eye out for is whether the rate of tasks started exceeds the rate of tasks completed. If you find that you’re starting tasks faster than you’re completing them, then you’ve clearly got a problem there.

Great. Other types of waste, hand-offs from one person to another. Again, when you see a hand-off from one person to another, again that should be something that you should be training yourself to spot as an insidious type of waste in a company. Hand-offs from one person to another, [Ellen 00:25:27] Ward says, “Hand-offs happen when we separate responsibility, so what we do, from knowledge, how we do it, from action, the person who’s actually doing it, and from feedback, the person who learns from the results.” Every time those things are separated, you typically have to add an extra person to the project and you have work handed-off from person to person. Or whenever any of the above are just arbitrarily split between two or more people. The reason it’s bad is because every time you have work handed over, firstly two people have to … no-one knows what’s actually happening anymore because everyone only gets part of the picture. And also the new person has to do loads just to bring themselves up to speed with what’s happening. So often you have to load two persons brains up with the information. And there are lots of other reasons.

An extreme example of this, I have a friend who works for a company where every month they were creating microfiches of all the documentation and posting them off to a different office as back-up. And he thought, I wonder what happens to those microfiches? So he phoned up the other office and said, “What do you do with them when you receive them?” And they said, “We actually forward those ones on to a different office in a different city.” And so he phoned up the person there and they said, “Oh well we forward them up to an office in a different country actually.” And finally he said, “Well what do you do with them when they arrive?” And they said, “Oh we don’t need those ones so they get shredded.” And it turned out at the end of this whole process that these microfiches were being shredded within about two months of them being created. And the reason was because no one person within that chain could see the big picture and could see how nonsensical it all was. So that’s a good example of one of the many forms of hand-off waste.

Other types of waste that you might not be aware of, over production, creating work that’s better than needed is very common and it’s often … you’ll see a company trying to perfect something much beyond the point where the customer actually cares. It’s often … there are certain things that the customers don’t care about. The next one, creating more work than can be used, and that’s where you’ll have someone who’s creating work that goes … a very common one actually is companies where there’s a large amount of regulation in the industry. So the web marketing team will create work but then the bottle neck is, oh, it just needs to go through compliance. And so there’s a limit to how much … unless more people are added to compliance or that bottle neck is removed then you end up with just more work than can be used. Next, creating features that the customers don’t value and I mentioned that before. Work that gets scrapped because it isn’t good enough, that’s another huge type of waste. An obvious solution to this is either retrain the person or remove them from the activity. If someone can’t produce the work to a decent standard, then there’s a good chance that you would actually save from them not doing it at all.

Right then, other types of waste, other types of waste. Overproduction and I’m not going into loads of detail here. Creating work that’s better than needed is an obvious type of overproduction. In particular beyond the level of polish that their client, or the end customers even care about. Another type of overproduction is creating more work than can be used. This is common in companies, in particular companies that are heavily regulated, where the bottle neck becomes the people who do compliance. And so that’s one thing to be aware of, that you’re not creating … well that there aren’t stages of your process that are creating work that can’t be used. Creating features that the customers don’t value. The next one, work that gets scrapped because it isn’t good enough. One option when that happens is to train the person and the other one is to remove them from the activity. If someone isn’t producing work that’s actually getting shipped, often you find that it’s not … you do need to question whether they’re actually delivering value at all if their work is not being shipped in the end.

Test it yourself, that’s one thing that’s very useful is make people responsible for their own testing. Facebook famously has a process where the engineers are responsible for what they call delivering the project all the way through until your Mum sees it. And so if you come up with an idea for Facebook you have to code it yourself and you have to push it all the way through and babysit it all the way through into production. So that’s like a very effective way of ensuring that the people who are creating the work aren’t just handing off stuff to people who are then fixing it for them. And the other one which sounds counter-intuitive but companies still … is lowering the standards. Sometimes one of the most valuable things we can do with clients is to … if they’re very hesitant to run tests, for example, on a site, or hesitant to send out surveys is to kind of encourage them to keep the throughput of ideas because the great thing about the web is that there is a control Z key and if something doesn’t work you can usually undo it. And so things don’t have to be so polished with the web as they do with print media for example, where the second the print run goes, that’s it. 10 million copies are printed.

And finally work spent checking work and work spent returning work is another type of waste that, once you start to become aware of it, you spot how much … you might spot that in your company you suddenly realize how much time is being spent on re-work.

A few tools for process improvement that we find really useful, the first is Kanban. If you don’t know what Kanban is it’s probably just worth looking it up on Wikipedia. Kanban is basically, effectively a task list where the number of tasks at each stage in your process aren’t allowed … you’re not allowed to have more than, say, three or four bits of work at any stage in your process. And what’s great about Kanban is the methodology prevents you from having too much work in progress at any one time, so it fixes all those problems I mentioned earlier. Another one is 5 Whys, it’s also called root cause analysis. Really root cause analysis is a better name for it except it doesn’t rhyme. 5 Whys is where whenever something happens or something goes wrong you ask why it went wrong, and keep asking why until you get to the root cause. Sometimes the root cause is a lot deeper than you might expect within a company. But once you’ve found the root cause, you often find that the root cause is the cause of other problems.

Deeper than that is the Theory of Constraints charts which help you to identify what the bottle necks are in your company and I’ll talk about that a little bit later. The Pareto principle is fantastic for helping you to prioritize work, especially if you ever work on, for example, prioritizing a call center, it’s amazing how using the Pareto principle you can … we’ve, in the past, reduced the number of incoming calls to a call center by, I think it was two thirds, by using the Pareto principle, finding the small number of problems that are causing the most outputs. Getting feedback from customers, and by customer I mean whatever is the next stage in the work flow. And that’s really valuable to ask. When you actually plot out what your work flows are in the company, even it’s okay then the person who designs the PowerPoint, the final designs then they are providing them to their customer who is the person upstream who’s implementing it. Whoever you’re providing work for, make sure that they give feedback so that you can understand exactly in what ways you’re providing value and in what ways you aren’t. Because often, again, the only real way of knowing how to improve the efficiency is to ask your immediate customer, for them to identify the waste that you’ve been creating or working on.

Now the next slide is not loading. Manuel, any ideas? Can you hear me still? So the sounds going but the slides aren’t moving. I think I might try … I can’t actually see the slides. I’m going to just start using … Thank you. Have you shared it with me yet? Got it, okay. I’m entering now. Got it, right. The Theory of Constraints. The slides moved back again. I’ll assume you can hear me and I’ll keep going.

Right, Theory of Constraints I want to tell you about because I think it’s a fascinating tool for understanding how … for seeing your business in a different light, in a more profitable light. I’m going to have to … Right I can’t see the slides which is obviously a little bit of an obstacle to me but I’m not going to let that deter me. I can see them on the … Manuel could you click on the slides for me, would that be okay? Right, if you could find the one that says Theory of Constraints one, because at the moment it’s on one that says three. Oh great, then I can actually see it on my own screen now, that’s even better.

Okay, Theory of Constraints was created by a guy called Eliyahu Goldratt and he was an engineer and he identified how we could use these techniques for … sorry the slides are moving around on my screen here. Right okay, it’s back onto that one. So I feel like I’m on some kind of TV game show challenge here. Right. Manuel I can just hear you speaking, it’s putting me off. Great, thanks. So what he did was he found this process which works well with processing plants but works well with businesses as well. And it’s finding the bottle neck in your business. Step one is identify the constraints, bottle neck’s probably a better word. Identify the bottle neck in your business. So how can you find out what the bottle necks are in your business? Here’s some little clues. Where is there work waiting to be processed? Where are those … before the bottle neck in a production plant you might find that before the machine that stamps the steel there might be a huge pile of steel waiting to be stamped because, obviously, work piles up before a bottle neck, it’s like a traffic jam. So where are the traffic jams in your company? Where’s the work piling up waiting to be processed.

Now next which resources are waiting for resources to send work to them? So which workers are waiting for other workers to send work to them? And at the end of each iteration ask the team, “What’s slowing you down or getting in the way of you doing a good job? What would help things to move faster, better or cheaper?” Now if you ask those questions you’ll probably find that you can identify where the bottle necks are. Now step two is exploit the bottle neck, so basically make sure it’s working flat out. This is subtle, it should be working flat out on the bottle neck scale, not doing non-constraint activities. So for example if you discover that your compliance … if you have a person who works on compliance and that’s the bottle neck to getting things out, the compliance person might say, “Well I’m busy. I’m working 12 hours a day.” But if you discover that the compliance person is spending six hours a day on compliance but last week they spent five hours fixing their computer and three hours showing a visitor round then you think, actually, that compliance person isn’t working flat out on compliance, they need to working flat out on that bottle neck scale. Someone else can fix their computer, someone else can do their expenses for them. Make sure that whatever that skill is they’re working on.

Now on the next slide is subordinate everything to the above decisions. Now ensure that the processes before are always ready to give work to it and that the process after are all ready to take work off it. This is subtle. To maximize throughput, freeing up the bottle neck isn’t just the most important thing in your company, it’s the only important thing. So if you have a series of people all waiting and the compliance person is the bottle neck, then you can think, oh yeah I get it, the compliance person’s the bottle neck, that’s really important. But actually it’s the only important thing. Getting that bottle neck to work faster. Because the over all flow rate of the whole process equals the flow rate of the bottle neck. So the only thing you can do to grow your company is to increase the rate through that bottle neck. Similarly, if someone or something is wasting the bottle necks time, they aren’t just an inconvenience, they’re slowing down the whole process. So I’d say this is slightly counterintuitive in that it’s not just important, it’s much more important than you realize because often the rate of growth of your whole business depends on the rate of production of that one bottle neck. Basically treat the bottle neck as a king, as we mentioned before.

Next, elevate the constraints, so widen the bottle neck. How can we increase the capacity of the bottle neck? Can it be more people? If that’s not possible can we reduce the requirements? Reduce what’s needed to be processed at that stage? Bearing in mind the huge cost. And finally sequential de-bottle necking. Once you’ve broken a constraint which means basically once it’s no longer the bottle neck, then there will be another bottle neck. So whenever you’ve finally said, “Okay then, compliance is no longer a problem. We made sure that the compliance person was working full time on compliance and we hired someone else to work in compliance, that’s no longer the bottle neck.” Then you’ll immediately go back and discover that somewhere else in your process, somewhere else in your work flows, there’s a new bottle neck that’s preventing the rate of growth of your company. So it sounds obvious but I think it can be easy to think, right then, at last we’ve tackled that problem and to give up, to move on, sorry, not give up.

The next one is good practices, small batches, work in small batches even when it seems like a bad idea. The reason to use small batches, the first one is what we always … a phrase we use a lot, seek the swift sword which is you want to know quickly whether it’s all going to work or not. So the quicker you can get something out the door, you learn the second whatever you’ve created is in the real world.

An example that I used to … when I was doing my PhD often I would have to photocopy 20 scientific papers and that usually meant going to the library, finding all the papers, they’re all in big volumes, having a huge pile of books heavier than I could carry, often. Take them to the photocopier and then photocopy those papers. And one thing I learned was that I should always start off my photocopying one paper and then going back and doing the others. Because always something went wrong. And the optimist in me always used to think, oh no this time it’ll be fine, I know the photocopying machine’s on. So I’d walk with all these papers and I’d get to the end, get to the photocopier and then I’d start to photocopy and I’d realize, oh, there’s a paper jam. It’s an engineers problem. Or the photocopier hasn’t warmed up yet or there’s a queue of five people waiting to use it. But it’s amazing how I could never predict, and it would always surprise me how many ways in which the whole thing wouldn’t work. And I eventually got this discipline of working in small batches because things often do go wrong and the best way is to work your way through the whole process just once.

The next one is, and you might think of this as tracer bullets. Often with web marketing it’s useful to think of tracer bullets just to get quick insights. Run small little mini projects, tiny little things that you implement just so you can learn exactly what’s going to happen. And then double down on something if it did work. You get the pay-off sooner which is great because work has a shelf life like I [inaudible 00:46:55] before. And less money gets tied up in inventory. Obviously if you work in small batches then the amount of work that’s all in your inventory is obviously less. Working in small batches is hard though. You need, in particular, to break your habit of bundling work. Don’t have nice-to-have features, just ship the minimal viable product, the minimum that’s needed to prove what it is that you’re about to try to prove or test. Craigslist, Amazon, Facebook, Google are great at this. They all avoid unnecessary graphics. They always just add the basic minimum of what’s needed. A couple of phrases that are useful is YAGNI which is you ain’t gonna need it. Often people find that programmers often anticipate features that’ll need to be in the software that never turned out to be needed. The other one is do the simplest thing that could possibly work, because it’s surprising how often it does work.

There’s a book called Obvious Adams which you can get online because it’s a really old one. I’d highly recommend it. It’s the shortest business book you’ve probably ever read. You can read it in about 15 minutes and it’s all about this guy Obvious Adams who’s always doing the simplest thing that could possibly work and getting away with it. He wants a job, there’s a large company so he starts off by contacting the founder of the company and saying, “I would like to work for you. I’ve no idea if I’ve got the qualifications yet but it just seemed like the obvious thing to at least let you know.” And yeah, it’s amazing how easy it is to ignore the easy option, the straightforward option.

Why it’s hard, combining one task with other tasks that need doing but aren’t necessarily dependent. So the most dangerous phrase in business is, “While we’re at it …” You don’t want to think while we’re doing X it’ll be more efficient if we do Y at the same time. Because what you’re doing there is creating a project that’s got dependencies, it’s probably twice as big. So whenever you think, oh well, it’d be crazy not to … we’re working on the checkout process, we’re going to revamp the whole checkout process, there’s something quite exciting about those big projects but there’s also something massively inefficient. We know a company that’s worked for … I think it took over three years to redesign the checkout process because they kept adding requirements into it. The other one is don’t polish one side of the bridge while the other side is rusty. It’s tempting to keep working on something until it’s perfect, but it’s better to just improve it until it’s no longer the worst feature of the business and then go and work on the worst feature of the business.

Second good habit, reuse or repurpose existing work. People tire of their own work long before others do. So spend more of your effort promoting work once it’s published. There’s a famous story of how Henry Ford walked into his advertising company and said, “Are we still running that advert?” He said, “I’m sick of seeing that poster” or whatever. I don’t know if it was a poster actually. He said, “I’m sick of seeing that as.” And they said, “Well actually we haven’t begun the campaign yet.” So he was sick of his own … because he was seeing these things internally all the time, he was actually tired of something that hadn’t even been published.

Next, good habit number three, constantly seek customer feedback. Not just for the end customer but for the customer of every process. And surveys and user tests are great. We do an amazing number of user tests for everything, even for our HR documents we do user tests because it’s the easiest way of getting good feedback.

The next … I want to just tell you something that we find really useful and we’ve not heard anywhere else apart from on a health and safety course from years ago. There’s a … I think it’s the DuPont health and safety process, I think it’s that one, which is a sequence of prioritization. Remove the activity. I’ll tell you an example actually. At university, when I was doing my PhD, I had to dissolve some glass off the surface of some ceramic. And removing silica is not easy because not many things dissolve glass. You can’t use washing up liquid. So the only thing that could be used is hydrofluoric acid. Now the problem with hydrofluoric acid is that you need … it’s really nasty stuff because it goes on your skin but it doesn’t burn your skin. It actually soaks into your skin and then dissolves the bone, it goes through to your bone and dissolves the bone. So you don’t immediately feel the problems but then, obviously, it’s agonizing if you get just a drop on your skin it would be really disastrous.

And so I found this process for health and safety is really useful and it’s applicable to all work. Step one is remove the activity, so I’m going to show on the right hand side the relevant things for hydrofluoric acid. So to remove the activity I could just abandon the experiment and think, do I really need to know what happens if I remove the silica off the surface of this? That’s the easy one, just don’t do it at all. The next one is okay then, if you’re still going to remove the glass, can you remove the hazard? Can you remove the hydrofluoric acid from your process somehow? So yeah, don’t use hydrofluoric acid, that would be one option, look at alternatives. Maybe I could have ground off the surface glass.

Next would be use an alternative to the hazard. And so for hydrofluoric acid it would be use a different acid. Would there be a way of doing that? Enclosing the hazard, so that would be use a fume cupboard, so at least it’s kept away from me or some other way of means so that the human doesn’t need to interact with it. Next, and usually this is almost always second from last, is wear protection and so that was one option, for me to wear one of those protective bibs with a big face mask and the kind of thing you can see [Walt 00:52:59] wearing there but with the face mask and proper gauntlets. Actually there’s normally someone standing over your shoulder with an antidote as well, so if you did get splashed then they can immediately step in and do it. And finally is follow a good procedure, which is like a safe system of work. A document called … effectively tips to handling hydrofluoric acid. Which is always the worst option because it’s prone to human error.

Now that’s how to handle hydrofluoric acid, but what we’ve found is that this same approach is useful for any kind of work. And so this is how it can apply to conversion. Because obviously everyone’s busy so it’s how can we get work done quickest? Step one, what would removing the activity be? For conversion it would be removing the activity, question whether we need to do this at all. One of our favorite phrases is the simplest type of code is no code at all. So, in other words, the easiest type of code to maintain is just not bothering with that. So really questioning whether the whole activity is needed. One of our clients, I remember one of the first things we did when we took on the client was we told them to put … it was something like 80% of all their projects on complete hold. We said, “Look, we don’t think those things are as important as this one.” And so they immediately reduced their workload massively but us pausing most of what they were working on. That’s an amazing way of getting things done. Obviously it requires having some insight into what’s fruitful and what’s not, but it’s very valuable.

The next one is removing the hazards. Okay we’re going to carry on with the activity, but how can we remove the problematic aspect of it? And that’s removing the bottle neck which I’ve mentioned before. The next one is use an alternative to the hazard. So in that case it would be, I guess, not just removing the bottle neck but using a different method for getting the same result. Is there a method that somehow bypasses that bottle neck? We don’t just have to widen the bottle neck but we’ll do it in a completely different way. Next is enclosing the hazard which, probably the nearest analogy to this is, okay we’re going to have the problem, let’s just handle it as well as we can which is delegating or automating the work. This is a really dangerous one to do before you’ve done steps one to three. It’s surprising how many companies delegate or outsource or automate work that they perhaps shouldn’t even be doing.

The next one is wear protection which is basically saying, okay we’re going to go ahead but let’s just have systems in place for avoiding the problematic aspect. And that would be just a more efficient workflow. And the final one is follow a good procedure which is just, yeah, having a good work flow, documenting the work flow. It’s the real low end … it’s effectively not far off let’s just try and be diligent and let’s put our heads down and do a good job. The key thing is so many people when they’re looking at their task lists start off at the bottom and then forget about the opportunities that you get at the top. There’s so much opportunity for identifying steps one and two in particular, how you can remove the activity and remove the bottle neck. And that’s really … when we look at the amazingly fast moving companies, that’s so often what they’re doing right. They’ve done an amazing job of saying no. If you’ve read the Steve Jobs book, Steve Jobs was amazing at strategic neglect and removing whole activities or whole departments or whole product ranges and simplifying to the point that by the time Apple was actually working on the lower half of this list, it was on stuff that was absolutely definitely worth doing. And so I’d say that’s a huge thing within your company to be looking at your workload in that way.

And that’s all I’m going to talk about for now. That’s probably enough work flows for now. I hope that was really useful to you. I’m sure it will be. A few things that you might want to do now, if your annual revenue is more than a million dollars and you’d be interested in speaking with us then go to contact us and we can discuss how we could help you to grow your business quickly and reliably. Next is if you want to work for us then there’s our careers page there, so get in touch. We’ve paused recruiting at the moment but we’ve got a capture form, so if you want to register interest than we can let you know when we’re taking on consultants again. Next if you want to learn more about conversion then we’ve got a great CRO toolkit at conversion-rate-experts.com/learning-zone which you’ll find really useful which contains landing pages of specific examples of pages that we’ve used to grow companies like Moz and Crazy Egg.

What you should do now

1. If you’d like us to work on your website—to dramatically improve your conversion rate and profits (like we did for all these companies), then claim your FREE website strategy session. On this free phone consultation, one of our experts will discuss your conversion goals and suggest strategies to double your sales.

2. If you’d like to learn conversion for free, go to our “Learning Zone” page, where you can download templates of million-dollar winning pages. Or, if you’d like us to build your company’s in-house capabilities (not for free), then contact us and we’ll discuss your requirements.

3. If you’d like to work for us—or see why our team members love working for us—then see our “Careers” pages.

4. If you enjoyed this article, then so will your friends, so why not share it on LinkedInTwitterFacebook and Email

All of our articles are subject to our Testimonial Protocol, which is described here.

Next article

An extra $300,000 for Piktochart, a B2B software company

Previous article

The fastest way to understand users: talk to a “VOC Aggregator”