In this episode of In-Ear Insights, Katie and Chris talk about code quality assurance, process management, and how often you should be updating your standard operating procedures. What happens if you don’t keep your processes and code in good working order? Tune in to find out!
Subscribe To This Show!
If you're not already subscribed to In-Ear Insights, get set up now!
- In-Ear Insights on Apple Podcasts
- In-Ear Insights on Google Podcasts
- In-Ear Insights on all other podcasting software
Advertisement: Data Science 101 for Marketers
Do you want to understand data science better as a marketer? Would you like to learn whether it’s the right choice for your career? Do you need to know how to manage data science employees and vendors? Take the Data Science 101 workshop from Trust Insights.
In this 90-minute on-demand workshop, learn what data science is, why it matters to marketers, and how to embark on your marketing data science journey. You’ll learn:
- How to build a KPI map
- How to analyze and explore Google Analytics data
- How to construct a valid hypothesis
- Basics of centrality, distribution, regression, and clustering
- Essential soft skills
- How to hire data science professionals or agencies
The course comes with the video, audio recording, PDF of the slides, automated transcript, example KPI map, and sample workbook with data.
Sponsor This Show!Are you struggling to reach the right audiences? Trust Insights offers sponsorships in our newsletters, podcasts, and media properties to help your brand be seen and heard by the right people. Our media properties reach almost 100,000 people every week, from the In Ear Insights podcast to the Almost Timely and In the Headlights newsletters. Reach out to us today to learn more.
Watch the video here:
Can’t see anything? Watch it on YouTube here.
Listen to the audio here:
- Need help with your company’s data and analytics? Let us know!
- Join our free Slack group for marketers interested in analytics!
What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.
Christopher Penn 0:02
This is In-Ear Insights, the Trust Insights podcast.
In this week’s in In-Ear Insights, it’s spring cleaning.
Specifically, I’m going back through and looking at some of my older code that I’m trying to modify to do a new thing.
So for those who don’t know, we have this process called most valuable pages where we try to take Google Analytics data, and evaluate which pages using machine learning are most impactful lead people towards conversion the most.
And one of the things that I heard recently from IBM actually was that they have a separate metric that they like called conversion efficiency, where you look at how many pageviews does it take before that page starts to generate some kind of impact.
And so I thought, it’d be great if we could take our output of most valuable pages output and bond that with IBM’s conversion efficiency, but we know that other versions of that are all last touching, you know, we have multi touch, thanks to machine learning.
So I opened up our code this morning, made a copy of it, because it’s never a bad idea to work on it never good idea to work on production code, and a break everything.
And as I was going through it, I was going, Oh, man, some of the stuff is really outdated as some of this, you know, it’s what the code I wrote is, this is probably almost two years old now.
And it’s, it’s almost embarrassing.
It’s like going back and looking at like, photos yourself, like college like, Oh, so Katie, how do we how do we prevent this happening? How do we reduce those embarrassing photo moments where you’re like, Ah, that’s not how you do that, or, I’m amazed this thing even works? Much less, you know, does what it’s supposed to do? Because, obviously, you want the, you know, the most efficient, cleanest code for any kind of process.
How, how do you read this the embarrassing moments here?
Katie Robbert 1:59
Um, well, I don’t think I have a lot of embarrassing moments from college.
First of all, I think any picture that I have from college is completely normal.
But maybe that’s just me, maybe I’m not embarrassed as easily.
But to your point, Chris, you know, one of the things that I see often happen, whether it be a software development project, or process development, or even just something as simple as how I pull a report, or the data in the report, what I don’t see often is that review or that post mortem, or that evaluation of, here’s what I did, now, can I do it better.
And so what often happens is that people will do something once and be like, okay, I did the thing, I’m going to move on with my life.
And that’s a symptom of us just being, you know, too busy or having, you know, being stretched and pulled in too many different directions.
You know, and so I understand why it happens.
But to your point, what the result of that the side effect of not going back and re reviewing things that you’ve created is that it’s out of date, it might break, and let’s say you are the one who created it, and then you left.
Nobody else knows how to fix it, or they would have to start from scratch.
And so the way to prevent this is to actually build into your project plan.
Now, when I say project plan, I don’t mean one specific project, I mean, sort of your year as a whole for your marketing team or for your development team.
There’s times when you’re not working on specific projects.
And that’s when you do those code reviews, those product reviews, those process reviews.
And so that’s a very long winded way of saying you need to build in those reviews on a more regular cadence.
Christopher Penn 3:47
So what’s the first step here? Because obviously, I’m halfway through this thing.
I’ve actually got a functioning prototype.
But what again, to your point, what tends to happen is once it’s functional, it’s okay, great, let’s move on to the next thing, particularly since this in its current incarnation is not yet paying client work.
It becomes part of paying client work once it works, and we change out or add to, you know, the analysis we do.
But right now, it’s still essentially unpaid work, given that a lot of companies are in is in a situation where they, you know, are demanding maximum productivity.
How do you how do you get stakeholders to buy into adding time to the process that is not profitable time.
Katie Robbert 4:34
You have to start small, you can’t just say, Alright, we’re going to dedicate 40 hours a month to unbuildable time, like that’s never going to go over well.
And so it needs to be in smaller increments.
And then you might say, Well, what can I accomplish in five minutes? What can I accomplish in 10 minutes? And so that does take a bit of planning on your part to sort of just outline what are the things that I have that I need to review and so maybe within that, Five minutes, you can at least make that list.
And then you can bring that to the stakeholders and say these are the things that haven’t been looked at in two years that we need to make sure are still up to date.
Cutting Edge won’t break are sort of the best practices.
Within that five minutes, you can see, you know, is there any kind of documentation that exists? You know, Chris, this is not a knock at you.
But you being more of that software development mindset.
My guess is that there’s no clear documentation.
So that if I had to go in and try to figure out what was going on, I’d be sort of raising my shoulders and shrug going, I don’t know, because he didn’t document anything.
And so there’s things that you can do to set yourself up for success when you do have that dedicated time.
And a lot of it is that prep work.
You know, maybe within like, maybe you have 15 minutes allocated well, within 15 minutes, can you start documenting what the process currently is, and then that way you stick like, you just sort of, you start to do as much as you can, within small chunks of time.
And then eventually, you know, you have a full, complete process that you can then sit back and go, Oh, okay, I see the five places in this process that could use improvement, and then you start improving them little by little, you know, by you, you start with step one, okay, so I used to do this, but now this technology has evolved.
So I want to do it this way instead.
And that actually works really well when you’re trying to do process improvement, because a lot like a B testing, if you do too much at once you don’t know what’s working or what’s not working.
And so really breaking it down into those milestones is gonna set you up for success in the long run.
And it becomes a lot less daunting, in terms of I have to do all of this work.
If you have, you know, think of it like, you know, let’s say you have to clean your whole house, or let’s say you have a mountain of laundry to do, that feels really daunting.
But if you start one room at a time, or one task at a time, and then you’re like, Okay, I got the bathroom clean.
Let me pause for a second and see what else I can do.
Okay, I got the kitchen clean.
Alright, let me pause for a second and see what else I have.
And just take it one piece at a time, it becomes a lot less overwhelming.
And the same is true for these evaluations that we’re talking about.
Christopher Penn 7:19
So as a practical example, I’m going to share my screen here, if you’re listening to this in audio player, just imagine a screen full of code with very, very few comments on it.
So Katie, looking at this just even rudimentary aliy.
This is just the part of the code that doesn’t clean up and assemble some of the measures for visualization.
From a process perspective, the only documentation on this entire page is that little blue line towards the bottom, which says make a report here.
What else? How would you start to pick this apart to say like, Okay, what is this exactly supposed to do? Because I can read it, and I have a pretty good idea of what it does.
And I actually ended up cleaning up a whole bunch of it this morning, because there’s some things that were really redundant.
But from that process improvement perspective, when you look at this, what would you what would the process be for picking this apart?
Katie Robbert 8:14
The first thing, the first step in the process would be to fire the engineer who wrote this, because it’s unacceptable.
Um, you know, if, if the only documentation someone has put in source code is make report here, that’s, that’s not going to fly.
That doesn’t help anybody.
And even Chris, as you’re describing it, you’re like, and I mostly know what it does.
You’re the person who wrote it.
And so that’s the problem.
You know, and I say this lightheartedly, because I know you and I know that you’ll be able to sort it out.
But what really needs to happen is there needs to be caught, you need to sort of break it up into sections, and have some sort of commentary within each section.
Now, there are, so you’re coding in our and so depending on the type of coding package that you’re using, there’s other ways to comment.
Um, but basically, what you want to do is chunk it out into sections to be like, you know, this first section, you know, you should have the following five things almost like your prerequisites, you know, you need to have a clean CSV file, or you need to have data from at least six months ago, those kinds of things.
So there needs to be some sort of, what do I do first, you can’t just open the code and run it, you must, obviously have some data to be running the code against you know, do you know if you’re doing, you know, software engineering on a website, like do you have to be running a certain kind of browser does it not work with like other browsers or operating systems, those kinds of things.
So it’s, you need to have those prerequisites stated upfront.
And then depending on how in depth your code is, you might want to break it out into different sections to say, you know, this section once it’s done, finish Sharing, cleans the code, cleans the data.
This section, once it’s done running, creates the analysis.
And this section, once it’s done running, actually generates the report.
And so that way, you know, if something’s not working, something breaks, you can actually go back and start to see where things are breaking.
So that’s the way I would break it out.
You know, I would try to avoid any sort of like shorthand or abbreviations, because what you understand might not be what a different engineer understands.
Christopher Penn 10:32
I’ve seen one thing I’ve seen that does help, at least me what I’m learning other people’s code, Dr.
Julia silicate, does screencasts where she essentially narrates, as she writes.
So she’s literally writing the code and then saying, you know, oh, we need to do this, we need to do this.
And even though she’s not telling you the name of the function, you can see what the function is on screen, as she’s describing it.
How does something like that, which, to me as a as a coder would be more efficient for documentation, play a role into improving the process?
Katie Robbert 11:04
It does, it definitely helps.
I think that that sort of brings you halfway in the equation, it really depends on the context.
And so if you’re using, you know, the screencast, as a training thing, then you do want to be calling out the things that you’re doing, because what ends up happening, is someone watching the screencast be like, I don’t know where you logged into, or I don’t know where you got that information, or, you know, they’ll probably have more questions.
And it will actually end up taking you more time.
But it definitely does help to sort of see what somebody is doing.
And as they’re explaining it, and why they’re doing it the way that they’re doing it.
You know, that’s kind of the context that’s missing from code a lot of times is, why is it being done this way? Or what is the outcome that you’re trying to achieve with this particular section, you know, you can usually make the assumptions of what the code overall is trying to do.
And so in this instance, it’s we’re generating a most valuable pages report.
So we know what the code is meant to do.
But within the code, the different sections like why did you choose this algorithm? or Why are you generating the code? And the analysis this way? Or, you know, if I’m the stakeholder and I say to you, I need to look at it this way.
Why isn’t it being done? Is that something that can’t be done?
Christopher Penn 12:23
Now, since a lot of folks probably don’t write code as part of their marketing jobs, how do you abstract this back to something more simple? Like, for example, putting together Google Data Studio report, does this same process apply to things that aren’t code?
Katie Robbert 12:40
And that’s when you’re talking about basically your standard operating procedures.
And so it’s almost kind of like a checklist of things that need to be done.
And so it can be as basic as step one, navigate to data studio.google.com, step two, open a fresh dashboard, step three, and sort of like listing out the steps because that’s essentially what code does.
Code is just a different kind of language that outlines step one, step two, step three, up until the output.
And so a standard operating procedure, which isn’t code is essentially doing the same thing.
It’s the set of steps that somebody does in a repeatable way to get to the same outcome time after time, after time, it’s a recipe.
You know, step one, preheat your oven.
Step two, mix all your wet ingredients.
Step three, mix all your dry ingredients, step four, mix them all together.
Step five, shove your shove all the cookies in your face.
Christopher Penn 13:43
standard operating procedures that are essentially code for humans.
Katie Robbert 13:47
Yeah, basically, I mean, it’s call it a recipe, call it a, you know, set of instructions, you know, whatever you want to call it, to make it something that, you know, you can wrap your head around, that’s essentially all it is, it’s a set of steps that somebody follows to get from start to end with a set outcome in mind.
So it’s like building furniture or building a house or even doing laundry, you can follow you could argue that there’s a procedure for that, step one, open the washing machine door, step two, put clothes in Step three, shut door, step four, add soap, you know, and then obviously, there’s variations depending on what type of washing machine you have, or the kind of laundry that you do and whether it’s darks or lights or delicates or you know, whatever.
But those variations are so small that you know, you’re the process is meant to be adaptable for any different kind of scenario.
But ultimately, the goal is to make it as efficient as possible to get from A to B.
Christopher Penn 14:50
How much detail is too much detail like if you’re reading out instructions for laundry, opening the washer door to me seems like too much detail.
Like, if you can’t figure that part out, you probably should not be using the washer like, you know, do not put gasoline in plastic shopping bags seems like a pretty straightforward thing that if you don’t know that, you probably shouldn’t be using a gas pump.
The same thing when I look at code, like there’s some things that I document extensively, like I Oh, this is the part that breaks like, if this isn’t right, this whole thing falls apart.
Other things like, you know, make bars blue, to me is like, does that need documentation?
Katie Robbert 15:28
So this is, this is where this actually starts to fall down is those assumptions, Chris.
And so what happens is you’re going into it assuming a certain level of knowledge.
And that’s actually the incorrect way to do it.
Because that’s where people start to get confused of like, Well, you didn’t tell me to make it blue.
So I’m not going to go ahead and make a blue, I’m just going to leave it whatever default color is supposed to be.
And then someone goes, Well, that’s wrong.
And so the way that you should be approaching it is the lowest common denominator of understanding of how to complete a task.
And so it might sound like, Oh, well, if you don’t know that you’re supposed to open the washing machine, before you put the clothes in, like, you’re making the assumption that someone has ever done laundry before in their life.
And so you want to be able to create a set of instructions for anybody to be successful, regardless of their skill set.
And so, you know, if someone has done laundry before, then they’ll probably skip the first few steps.
Or they’ll just use the standard operating procedure as a guideline to say, Okay, I genuinely know how to do laundry, let me just double check and make sure that I’m doing it the correct way.
So same thing with code.
If somebody doesn’t know, to first open up the system, then, you know, they probably never done it before.
But they’re like, okay, I’ve coded before, I know which system to go to, I know, the different commands to navigate, I just need sort of that gut check of the thing.
And so you can’t make the assumption that everybody knows what you know.
And that’s where process development tends to fall down and not be successful.
Is that, that thinking that well, that’s obvious.
So I don’t even have to say it.
You’d be surprised how not obvious it is the majority of people?
Christopher Penn 17:15
Then how do you how do you balance that with the fact that if you had to document every single line in your code, your development process would take 10 times as long? Because obviously, you know, we have not done that ourselves for that very reason, because documenting every line of this, you know, 2000 line program would take forever?
Katie Robbert 17:34
Well, you know, it’s it’s all relative.
And so it comes down to, you know, what is it that you want to accomplish? Will you, you know, will you forever be the only person in the company who codes or do you want to scale the team to have other people coding alongside you.
And so that’s where you start to strike the balance is figuring out what’s the business goal.
And so if your goal is to hire people who have 10 years of coding experience, for example, you probably don’t have to document every single line.
But if your goal is to create an internship program, where you’re starting people fresh right out of college, or haven’t even taken a coding course, then you probably want to document every single step, because they won’t know it otherwise.
And so it really comes down to you deciding upfront, what you’re going to be doing with these processes.
And so let’s say let’s go back to the laundry example, let’s say you’re trying to teach your children how to do laundry for the first time, you probably want to be overly overly detailed, so that they don’t, you know, put bubble bath into your washing machine and then flood your whole house with soap.
versus if you’re just leaving instructions for you know, your teenager who’s done laundry a few times, it’s more of a reminder system of you’re not saying like open washer, you’re just saying, you know, use the tide for this or use the game for this or, you know, friendly reminder, don’t mix your red shirt with your white underwear kind of thing.
Christopher Penn 19:06
Yeah, like my wife will say just, these are the three settings you need to remember when stuff when you have to put the laundry through.
It’s like, okay, I can remember that part.
Katie Robbert 19:13
Christopher Penn 19:15
So what if we don’t have a specified goal? Like, what if we don’t know, like, in this case, your point, are we if we hire somebody on, there’s pretty good chance we try to hire somebody who’s smarter and better than me at coding, right? So a lot of those assumptions wouldn’t necessarily be in there.
So and I don’t know that we would be having interns per se.
So when I look at this code, right now, you know, for the next year, at least, it’s probably still just going to be me tinkering around in here.
So at what level should I be building documentation? Should it be for that like business continuity, hate somebody got hit by a bus kind of thing? Or how do you how do you make those decisions?
Katie Robbert 19:56
Well, I mean, it sounds like you’ve already started in your mind.
Making some of those decisions.
And so what you’re saying to me, what I’m hearing is that it’s not likely that we’re going to be starting fresh with interns.
So that like, very specific level of detail is probably not needed.
It’s more of a business continuity thing, where if for some reason you can no longer create the code or maintain the code, we need somebody who does know how to code to be able to pick it up where you left off.
And so that can be right, there is a decision, that’s a starting place.
Christopher Penn 20:26
And with standard operating procedures, things like Google Analytics and stuff is that does it the same or is it because there’s more humans involved from the beginning, you have to tune it to a different level.
Katie Robbert 20:38
I personally always tune in, again, to that lowest common denominator of understanding, I go in with the assumption that nobody has ever seen this thing before.
And that it’s brand new, and it’s scary.
And it’s, you know, something that, you know, is really, really important that somebody needs to learn how to do.
And so that’s personally how I like to approach creating those instructions.
And then once you get people using the instructions, you get that feedback of, Okay, these are the five things that we’re missing, or it’s way too detailed, and it takes me too long, you know, so then you can start to revise those processes.
Again, it’s like anything else, once you create the process, it’s not a one and done, you need to build in that time to review it.
Because, you know, maybe it’s changed maybe the way that you access Google Analytics is different from how you used to do it two years ago.
And so your process is out of date, or maybe the team that you felt, is so skilled to Google Analytics, that the first 10 steps actually just slow them down, because they’re following the process.
And so you need to reevaluate that documentation, whether it be in your code, or it’s a standard operating procedure for something, you know, on a regular basis, at least once a year, at least once a year, you need to carve out some time to do it.
You know, let’s say, Chris, you buy a new washing machine? Well, those three settings that you need to remember are probably going to be different.
Christopher Penn 22:06
And who doesn’t like having foam foamy, so well over the floor of their basement?
Katie Robbert 22:12
I definitely don’t enjoy it, especially since my washers on the second floor.
Christopher Penn 22:20
In the work that you’ve been doing you so you’ve been doing a lot of change management within agencies helping agency teams, you know, sort of get their stuff together.
Are you seeing? Are you seeing similar? I guess things where there’s a bunch of processes that don’t really aren’t really written down? And if so, how would you compare contrast that with, you know, my poorly documented code?
Katie Robbert 22:44
What I’m actually seeing is there’s a lot of templates, but no instructions on how to use the template vs.
code with no documentation, basically, yeah.
And so I think that there’s a misunderstanding that if you have a template, you’re done, that’s not true.
Because someone still needs to know how to use the template, someone needs to have an understanding of roughly how long should this task take me.
And I think that one of the things that’s almost never built in is some sort of a measurement of efficiency.
And so taking that baseline of, well, how long does it take you to code now? And if you build in process and start using that process, how long does it take you? Is it slowing you down? Is it speeding up? Because if you’re trying to revise the process, you need to know, you know, am I adding more to it? If I’m adding more to it, is there more value? Or am I cutting things just to make it go faster? You know, and so there’s no measurement, there’s usually no steps or instructions, there’s a lot of times just a template, and the template is not something that everybody is consistently using.
And so, you know, the other thing that I’m seeing is that lack of consistency.
Again, it’s that one and done mentality, a lot of people bring in a change management consultant to bark out a bunch of instructions of here’s all the things you need to change, go ahead and change them.
People change them, the consultant leaves, and then they’re like, Okay, well, we did that.
What, huh? There’s no follow through, there’s no consistency.
And it’s something that you need to continue to, you know, I guess, remind people like, Hey, you know, I saw that you created a blog post, but you didn’t use the template, here’s the template, you need to be using the template.
And it does feel like I can’t stand being like constricted to these things, these templates, these boxes, these processes.
But what ends up happening when you do that, is it becomes almost like muscle memory.
And then you free up all this other time to be creative and innovative.
And that’s Chris, you know, to the first question you’re asking is, that’s where you’re able to carve out that time to innovate and update and change and evolve, is because all of those other things just become, you know, a well oiled machine.
And that’s the goal is to get all of those other tasks.
Ron, that you build in time for yourself to do that non billable stuff, because the billable stuff becomes, you know, so second nature.
Christopher Penn 25:08
So the moral of the story then is to, to provide that additional context and to make the processes have processes of their own as long as it makes sense to do so.
I guess have my work cut out for me.
If you have if you have questions about anything we’ve talked about in today’s episode, hop on over to the analytics for marketers slack group, if it’s a free discussion group over 1700 markers over Trust insights.ai slash analytics for marketers and if you want to watch or listen to the show in some other format or channel, head on over to Trust insights.ai slash ti podcast where you can find the show in most places, most places not not all, but most places.
Thanks for tuning in, and we’ll talk to you next time.
Take care need help making your marketing platforms processes and people work smarter.
Visit Trust insights.ai today and learn how we can help you deliver more impact
Need help with your marketing data and analytics?
You might also enjoy:
Get unique data, analysis, and perspectives on analytics, insights, machine learning, marketing, and AI in the weekly Trust Insights newsletter, Data in the Headlights. Subscribe now for free; new issues every Wednesday!
Want to learn more about data, analytics, and insights? Subscribe to In-Ear Insights, the Trust Insights podcast, with new 10-minute or less episodes every week.