I’m sure it’s happened to you, maybe even where you work now: you wonder why in the heck you’re building a particular feature. It seems to you, and maybe to others you work with, that the company’s time and money would more profitably be spent building a different feature.
You might be right. I hope your company gives good avenues for providing that kind of feedback. But if it doesn’t, or if it does but your appeals failed, then get on with making your assigned features real.
That’s where strong execution begins — when the dust settles, even if you don’t fully agree, you get busy working the plan.
It’s obviously best when your company has brilliant ideas that you then execute strongly. But if you bring strong execution to even mediocre ideas, success usually follows. It’s a great overlooked secret: there is no substitute for everyone pulling the cart hard in the same direction.
Companies with brilliant ideas but poor execution fail. Or so, at least, it has gone at the companies for which I have worked. I was once in a startup with an idea that I thought was brilliant. But there was so much infighting inside that company that neither sales nor engineering could build a strong execution rhythm. That company is still limping along but to my mind its product never reached its potential.
If your company uses agile processes, you’ve hopefully broken your work down into small enough chunks that even if you choose a wrong feature to build, you won’t pull the cart in the wrong direction for too long. You’ll deliver a little software, see how the market responds, adjust your direction as needed, and repeat (and repeat, and repeat). Ideally, this system self-corrects.
By all means, if you disagree with the features you’re assigned to build, provide that feedback using whatever mechanisms your company offers. But after the plan is set, execute it as strongly as you can.
Did you know that a fly doesn’t scale? If you shot a growth ray at a fly until it was the size of your car, that fly would crush under its own weight. Its exoskeleton can handle only so much mass.
It’s a good thing flies stay as small as they are. But if you’re at a small, startuppy company, of course you hope your company grows. It’s what we all want, right? But meanwhile, there can be so much goodness, so much energy in a young company. You can wear so many hats! You can shoot from the hip! The hierarchy can be so flat! Everybody fits into one room! Communication is so easy! Everything and anything feels possible!
But much like the fly and its exoskeleton, your company’s light, informal structure will keep it from scaling. One day everybody won’t fit into one room anymore. You won’t be able to know everything that’s happening just by being in the room and on Slack. Stuff that used to happen organically just doesn’t happen well or at all.
That’s when meetings may creep in where there had been none. When the first process steps may start to be added. When software tools may augment or even replace simple face-to-face communication. When company-wide fun events may start to drop off because it’s too hard and expensive to include everyone.
The place starts to feel …corporate.
People sometimes fetishize startup smallness. It feels so good! But clinging to it will limit your growth trajectory. Like the overgrown fly, lack of process will crush your company. You need to change your ways of working to fit the company’s size. But this doesn’t have to be terrible. It totally can be terrible, if you do it wrong. Even if you do it right, it will change and even get rid of some of your company’s original goodness. But it enables new levels of goodness that you can’t imagine yet.
The tax of size simply must be paid. But don’t pay one penny more than you need to. At every step, add just the minimum process to keep things running smoothly. If you’re an individual contributor and thus probably not making the process decisions, perhaps you can (kindly, respectfully) ask if you can offer feedback, and if so, suggest changes to size the process to where you are now.
This is a bumpy road, and there’s no map. You have to feel your way through. But here are some questions you can ask that might help you.
What’s the least possible amount of process we can add to solve our current problems?
How can we break our organization into smaller units to keep the goodness of small?
How can we put light communication, visibility, and accountability systems in place that help the small units stay connected and deliver strongly for each other and for the business?
To the extent you’re able to successfully evolve your company’s exoskeleton of process and culture, you’ll smoothly adapt to the pressures that come with growth.
My colleague Matt Block recently posted on his blog a link to an article about how a software shop’s business model affects how well agile scrum works for them. It breaks business models down into emergent, essentially meaning that the company builds product to meet goals such as selling ads or driving traffic,and convergent, essentially meaning that the company builds product that directly serves a target market. The article argues that agile is made for emergent and is a poor fit for convergent. That’s just a sketch of the article; go read it to get the full flavor.
I’ve always worked for companies following convergent business models. We’ve made our money by selling the software we created, which made it always important to deliver a certain scope by a certain time. When those companies implemented agile scrum, they could never fully adapt a key principle of it: when it’s time to ship, you ship whatever is built. In a convergent world, scope is king; you ship when everything specified is built.
I e-mailed my brother, Rick Grey, a link to this article. It’s great to have a brother who does the same thing I do for a living as we can talk endlessly about it. I thought we’d have a conversation about how to scope an agile project, but instead he had a brilliant insight: What if agile is good for convergent-model companies because it tells you sooner how much your project is off track? He gave me permission to share his e-mailed reply, which I’ve edited.
– – –
What if the companies we’ve worked for and all the other convergent-model teams of the world are doing agile just fine? By “just fine” I mean “as good as they do waterfall,” which may not be “just fine,” but we’ll get to that in a minute. Meanwhile, consider:
Long waterfall project:
No one pays real attention to progress (there’s always next month to catch up)
Engineers go dark, checking out huge sections of the codebase and not merging them back for long periods
Engineers (who are notoriously poor estimators) claim 50% done when it’s really about 25% – and then, as the code-complete milestone nears, they (usually innocently) claim 90% done when it’s really 70%
A couple of days before the code-complete milestone, engineering finally acknowledges they won’t hit the milestone and delays delivery to QA – “but we’re 95% done, for sure”
Under the pressure of already having missed a deadline, developers quietly take shortcuts to make it possible to hit the new QA delivery date
Weeks and months of unmerged changes come crashing in, creating conflicts and compile/deploy problems, further delaying delivery to QA
QA, now staring with a multiple-week handicap on an already-too-aggressive schedule, quietly takes its own shortcuts
QA finds hairy showstopper bugs and so the ship date gets moved
Management is livid, so QA goes into confirmatory testing mode just to get it out the door
Agile project of the same size:
Much of the above happens at a smaller scale, one iteration at a time
You fail to deliver everything planned starting with the first sprint
Instead of spending 80% of the project thinking you don’t suck as an organization and the last 20% realizing that you do, agile lets you feel like you suck every step of the way
Takeaway for management: “agile sucks” and/or “we suck at agile”
I assert that most teams are bad at delivering under a convergent business model. The hallmark pathologies of software delivery under a convergent model are too numerous and powerful for most teams to overcome, but their struggles are masked by waterfall until the end. Agile surfaces the problems every iteration. You feel like a loser by week 4 instead of week 40.
But this is actually a win. You get better project visibility and a tighter feedback loop, meaning you’ve got a better chance to make adjustments earlier to get the most out of your team you have. Embrace the feedback loop as a chance to make things better, and learn not to view it as proof of how much you (collectively) suck.
– – –
I will add that agile also helps you keep resetting expectations within your organization, because it makes it standard practice to keep reestimating what it will take to finish everything. This is just what I was talking about in my last post (read it here).
I’ve said it to my test teams many times: Making software isn’t quite engineering. Building a bridge – now that’s engineering. You determine how long the bridge needs to be, how much load it needs to carry, and what kind of bridge to build (steel truss, concrete arch, etc.), and from there it’s mostly mathematics and physics. Just run the calculations and you’re good.
We have bridge-building down. With a couple of notable exceptions, such as the Tacoma Narrows bridge which heaved and twisted and finally collapsed (video here), new bridges seldom fail. Old bridges fail sometimes, but it’s reliably due to accident or neglect.
My apologies to any civil engineers who stumble upon this post. I’m sure you’re cringing that I’m overlooking many subtleties of your discipline.
There’s nothing subtle, however, about how often software fails. Our users aren’t happy about it, but they aren’t surprised by it, either.
For any thing you ask a software developer to build, there will be a whole bunch of valid ways to do it, each with its own unique ways of creating failures. This is especially true when when that developer enhances existing software that he or she didn’t make in the first place. It’s tough to predict exactly how the enhancements will affect the rest of the software. The more lines of legacy code, the more time and analysis it takes to think that through.
If a developer had unlimited time and money, it might be possible to deliver perfect software. Ah, a developer can dream! But here’s where bridge-building and making software have an important thing in common: time and money are never unlimited.
I sympathize with the folks who call software a craft. People who make software use tools and knowledge in its design and construction. These are hallmarks of craft.
Another way that software is like craft is that it’s difficult to fully separate the design from the making. Even when one person designs the software and another writes the code, the coder has to make a bunch of lower-level design decisions along the way.
The software craftsmanship movement meets corporate resistance because revenue and profit ride on what we build. Our companies need to sell features to meet revenue projections, or deliver bug fixes to retain customers. That’s why timed delivery is so important: if you wait too long to deliver, the opportunity to grow or retain revenue begins to shrink.
Feeling pressure to deliver, yet knowing that if we deliver junk we’ll be in an even worse pickle, we tend to manage software-development projects like engineering projects. I think we feel like we have better control when we manage them that way. But that feeling of control can’t mask it: no matter how tightly you plan a software project, no matter how you shape your development and delivery processes to mitigate risk, no matter how much you try to predict the troubles you’ll encounter, you will discover things along the way can seriously derail those plans. It happens in two-week scrum sprints just as it does in ten-month waterfall projects. Discovery is simply endemic to software development.
As a software project manager, I try to build in buffers for the unknown. I also steer projects daily based on what we discover, adjusting plans and communicating impacts to whomever needs to know. I try to make sure our development practices deliver the best possible code to test, and then I try to arrange testing to find the worst bugs first so that near the hoped-for end, only minor bugs remain. Despite all that, important bugs still sometimes reach the user.
We ship when the software is good enough. What “good enough” means varies from context to context, but it is unfailingly short of perfect. Shipping at good enough means you succeeded.
If I delivered bridges that way, I’d never drive over one I built.
“Our product is a Web app,” the CEO said. “Why wouldn’t it work flawlessly on an iPad?”
The answer, of course, was that we didn’t build it with the iPad in mind. When we got it to work on our PCs on Chrome, we declared it done. But some guy at one customer site decided to run it on his iPad, and he was upset because it didn’t work “at all.” Support escalated the case to Engineering, where one of the testers poked around the product on her personal iPad. She found that she could log in, navigate, and enter data on any screen. “It was a pain. I’d rather do it on my laptop any day. But it did work,” she said.
So we were puzzled by our user’s bad experience. And then we learned that he had an original iPad. Our tester had a fourth-generation iPad – on more modern hardware and the latest version of iOS. Unless we wanted to shop for an old iPad on eBay, we were never going to get to the bottom of this complaint. The user wasn’t thrilled when we told him that we couldn’t support his old tablet.
Crushed under the weight of platform support
If you don’t tell your users what they can use to access and run your product, they will run it on whatever they want. When it doesn’t work, they’ll demand you support it. This will crush you. You need to set some limits.
This isn’t meant to be done by fiat. You need to think it through: What are your customers likely to want to use? You need to be there. How will they access your product – devices, operating systems, browsers? If your product will be installed on their equipment, what hardware (or virtual servers), operating systems, and databases are they likely to use? If your product interfaces with other products, what versions will your users want to use?
What you’re doing is limiting the edge cases, like that user’s original iPad. You’re also limiting the list to what you can meaningfully test.
Creating a matrix of supported platforms
Create a spreadsheet that says your product runs on and with these versions of these other products. If you support several versions of your product, the list of supported platforms will probably vary a little version to version.
Give the spreadsheet to support and sales and say, “Use this to set expectations. If they want to use something that’s not on this list, try to talk them into sticking to the matrix. If they insist on going their own way, tell them we might not help them if they have problems.”
This isn’t a once-and-done job. New versions of your supported platforms will be released. Your users will want to use entirely new products with your product. You’ll have to decide when and whether to add them to your matrix. You’ll need to remove old platforms from the matrix sometimes. You need to keep gathering information and keep updating the matrix.
You can’t support it all: why I hate Android
Still stinging from the customer who had the bad iPad experience, the CEO asked, “Do we have any idea how the mobile Web site behaves on Android? It needs to run on Android, too. It needs to run on any device they have in their hands.”
The problem with Android is that there are eleventeen base versions in active use, each of which a vendor can customize for their devices. And the devices themselves run the processing-power and screen-resolution gamut. When I explained that I’d need to buy a hundred devices and hire twenty testers to test them all, the CEO became less excited about Android.
You can limit the scope of Android, and any other platform with a lot of permutations, by deciding which permutations you’ll support and testing only those. As I write this, the 2.3.x (Gingerbread) versions of Android are the most widely used. You can buy a couple of popular devices on that version and test just those.
The only challenge is that where there are lots of permutations, you’ll have plenty of users on unsupported permutations who aren’t going to be happy.
This reminds me of the browser interoperability problems from several years ago, and why companies usually limited their Web apps to work only on Internet Explorer. Browsers are better now, and it’s more common for Web apps to work on any modern browser. Here’s hoping Android settles down as browsers have.
You can’t test it all: Supported vs. Certified
The items on your supported platforms matrix needs to be things with which your product actually works. As much as possible, design your product up front to work with them. But if you’ve just created a supported platforms matrix for a product that’s been around a while, the best you can do is test your product with those supported platforms to find out where the bugs are. The knee-jerk approach is to do a “full” (whatever that means) regression test on every platform.
One place I worked, our supported platforms matrix grew huge. When we printed it, we needed a jeweler’s loupe to read it. For each version of our product, there were fifty or more supported platforms. I couldn’t afford the army of testers I’d need to do “full” regression passes on them all.
But all that testing really wasn’t necessary. You can look at what’s new in the latest version of a supported platform and make a pretty good assessment of the risk to your product, and size your testing accordingly. For example, a product I used to work on ran on the Oracle and SQL Server databases. When Oracle 11g Release 2 (220.127.116.11) came out, it offered enough new stuff that we weren’t sure how it would run with our product. So in our next release, we did our normal release testing, including regression, on that database. We also ran some special tests that exercised our most complicated SQL queries on it. We found and fixed some bugs that way.
When Oracle issued 18.104.22.168, we looked at what was new in that release and saw no reason to believe that any of it would create new bugs in our product. Our experience was that if a database version was going to throw bugs in our product, those bugs would be foundational and would appear right away under even light testing. We had a bug-fix release coming up, so we installed 22.214.171.124 on that test server and tested the bug fixes on it. The release sailed through testing, so we shipped it and added 126.96.36.199 to the supported platforms matrix.
Platforms that got the so-called “full” test treatment were listed as “Certified” on the matrix. What we told our customers was that we had “performed a certification test” on the platform, had fixed the bugs we found, and were highly confident of a good experience on that platform.
Platforms that got light testing were listed as “Supported” on the matrix. Sometimes our analysis showed low enough risk that we did no testing and still called the platform Supported. We told our customers that we expected that they would have a good experience on that platform.
The bottom line for both Certified and Supported was that if the customer experienced a problem, we’d take their call and resolve their issue.
The bigger bottom line is that you want a manageable list of platforms with and on which you believe your users will have good experiences using your product, a list that won’t kill you to keep up with. Through understanding what your customers want, limiting the list to the most common and important platforms, and varying the depth of testing based on risk, you can keep up.
I have been astonished in my life by how few problems are truly unsolvable. I have also noticed that, most of the time, when a problem ends up not being solved it is for one of two reasons: people deny the problem, or they won’t work together on the solution.
There are any number of ways to make software. All of them involve programming, of course; it’s the core activity. But there are other jobs to perform, and the bigger the software being developed, the more people you need to perform them. It becomes necessary to organize the people and form work processes. The ways to do that range from loose and chaotic to structured but flexible to ponderous and formal. They all have their strengths and weaknesses; they all can work. I’ve made software in all of these situations.
One of my past employers sequestered its programmers in a room to hack code together as fast as they could. When they were done, a couple people would play with the product for a week or two in hopes of finding some bugs, which the programmers might or might not fix. Then we shipped it to five customers, who upon installing it invariably found that it failed spectacularly. When we fixed all the problems they found, we chose five more customers. We repeated this process until the software stopped failing, and then announced to the rest of our customers that it was ready.
At the other end of the scale, another past employer had such a heavy software development process that following it felt like slipping into a straitjacket. There were reams of documentation to write and keep with layers of approval at each step. I led the test team. We had to take a screen shot of every test step, print it, and put it in a folder. At the end of a project, we had boxes and boxes and boxes of printouts, hundreds of pounds worth, that we would send to an offsite storage facility. This was in case our only customer, which happened to be the U.S. government, wanted to audit us.
Both companies are still in business.
A third past employer carefully hired smart people and trusted them to know what to do. Because they hired well, this worked for a long time. But as the company grew, this approach became more and more difficult to manage. Product quality became a problem. A big part of the software was an accounting package, and in one fateful release the general ledger simply would not balance. This affected every customer, and to make a long story short, it took the company almost a year to fix it. Several customers quit us.
Too often it takes great pain to drive important change. This pain made us face that we needed more structure to deliver successfully. So we hired a consultant to guide us toward a better software development process. He showed us a way with just enough checks and balances that we could have greater confidence in our work without being too bogged down. We also hired a consultant to help the management team (of which I was a part) work more collaboratively. She taught us how to lead the company through this transition.
It worked. In the very first product release under the new process, quality problems fell off dramatically and we delivered on time for the first time anybody could remember.
The more important hire was the second consultant. The new methodology was good, but it wasn’t magic, and it might not have worked for us if we had not done a good job helping the team through the changes. You see, a few people didn’t understand why the old way couldn’t still work and still others thought our new process wasn’t going far enough. Our task was to overcome the resistance and get everybody singing from the same sheet of music, and it was the hardest thing we did. But we pulled it off.
It’s a rare organization that faces its failings and works to change. It’s a rarer organization still that drives change collaboratively.