Episode 34: The Contraindications of Key Performance Indicators with Keith Perhac

Keith is the founder of SegMetrics, helping founders understand their true lead value using a data backed approach.

On Air, Keith and Rob discuss why your KPI's might be setting up blinders as your create a marketing strategy for your businesses, how outliers can create a new train of thought in your marketing plan, and how to identifying strategies that can truly move the needle in testing conversions.

Keith Perhac has delivered a few talks at MicroConf, from his first Attendee Talk back in 2015 to his keynote presentation at MicroConf Starter in 2017. You can check out his talks here: https://microconf.com/speakers/keith-perhac

MicroConf Connect ➡️ http://microconfconnect.com
Twitter ➡️  https://twitter.com/MicroConf
E-mail ➡️ support@microconf.com

MicroConf 2020 Headline Partners

Stripe
https://stripe.com
Twitter ➡️ https://twitter.com/Stripe

Basecamp
https://basecamp.com
Twitter ➡️ https://twitter.com/Basecamp

TRANSCRIPT

Rob Walling: [00:00:00] And we are live. Welcome to MicroConf On Air.  Every Wednesday we livestream for about 30 minutes and we cover topics related to building and growing ambitious SaaS startups. These are startups that don't need to work 80 hour work weeks, raise millions in venture capital, or drive us to the brink of burnout.

In this show, we seek freedom, purpose, and relationships, and we want to maintain those values while building interesting companies that can have an impact on the world, or maybe just our little corner of it.  So thank you so much for joining me again this week, whether you're joining us live and maybe you're in MicroConf Connect, and you're going to be asking some questions of our guest today, or even if you're listening asynchronously on the podcast feed that's microconfpodcast.com or search for MicroConf On Air in any podcatcher that you use.

Today I have the pleasure of, uh, bringing a Keith Perhac on the show. You may know Keith, as the founder of SegMetrics. Yeah. And he is , hailing from the Portland area. So it's a little smoky outside his house right now. But  Segmetrics.io If you want to check it out.  They're a team of about 10 people in the growth stage.  I was talking to him  beforehand about how can we give people a sense of, of where the company is? , and he said, "Look, we have product market fit and we're in the growth stages. We're starting to grow quickly. "

 If you go to SegMetrics.io, you can see their H1.  In essence, SegMetrics is a tool that helps you get 100% clarity on where your leads come from, how they behave and how much your marketing is really worth.  It allows you to get a handle on the KPIs that matter most for your marketing funnels is built by marketers for marketers.

And today, Keith and I are going to be talking about how KPIs can be misleading when optimizing your marketing. So Keith Perhac. Welcome to MicroConf On Air.

Keith Perhac, Segmetrics: [00:01:43] Hey, Hey Rob. Thanks for having me.

Rob Walling: [00:01:46] Absolutely, man, um, that sky's look kind of clear from what I can see reflecting in your window as the air quality improved up there.

Keith Perhac, Segmetrics: [00:01:54] It's not bad right now. It was really bad last night. It was kind of bad this morning. Uh, we'll see. It's really weird, you know, as the sun starts to set at and at like three o'clock it looks like an apocalypse out there at two o'clock. It looks like it's nighttime almost. So knock on wood that it's getting better.

Rob Walling: [00:02:11] Yeah for future. If you're listening to this in a year, this, these are the great, uh, fires, the Pacific Northwest and California happening and fall of 2020. So if you are in MicroConf Connect to  go to the #microconf-on-air channel. And you can ask questions of Keith or myself, and we of course will tackle those live.

But Keith, we're going to talk about how KPIs can be misleading  when you're optimizing your marketing, I'm going to define KPIs as , key performance indicators for folks who aren't fluent in an MBA speak. So I don't, I don't like KPIs.  [indescipherable].

Key Performance Indicator.

I mean, they're, they're critical and they're needed in larger orbs. It's just the stuff. As a startup founder, I'm always like get out of my way. I want to know process, you know, um, key performance indicators are the critical indicators of progress towards an intended result. KPIs provide focus for strategic and operational improvement.

They create an analyst basis for decision making and help focus attention on what matters most. And that is a definition I took from something on. Google. So its probably the little, the little inset with the questions, but to me, I've realized that it's like once Drip got acquired, went from a 10 person team up to a hundred person team, basically.  it's like each team needs something to focus on  and if it's numeric all the better. For customer success team, maybe it's it's NPS, or maybe it's churn, or maybe it's trial to paid conversion rate. And that's the, those are the key performance they might be looking at it.

Sales, obviously it's going to be closed deals. Dollar amount of deals closed, MRR growth from sales. For engineering, those ones are always hard, right?  The CEO, when I left, was like, how can we measure engineering? And I'm like, well, it's not bug count and it's not number of features shipped. It's not lines of code shipped.

So I'm very skeptical of, of KPIs with development teams. But that gives you an idea.  If you're listening out there and , you've heard KPI's, or you really didn't know what they are,  they're just numbers that help you kind of guide your business right. And to drive it forward.

Keith Perhac, Segmetrics: [00:04:04] Yeah, I think that's, that's fully accurate. The one thing I would add on there is that the idea of KPIs is kind of like a scorecard . You don't want to have so many that you're just looking at a line of numbers.

But at "some point you want to know, "Are we doing better or are we doing worse than we were last week or last month? Or last year?" And as SaaS founders, most of us are looking at MRR. That's our top key KPI. It's like, what's our MRR? Is it going up or is it going down?

Rob Walling: [00:04:30] Yep. That was my top line KPI  too , when I was all my, when I was running all my SaaS companies and everything else, but then everything builds to that because then I would look at how many trials did we get this month? What is our trial to paid conversion rate? What is our churn? What is our net churn? What you know, and all those things contributed, but you're right. That number that it builds to is MRR it's growth rate.

  Cool. So, um, first question of the day.

All right. What are the main KPIs that people should be looking at in their marketing?

Keith Perhac, Segmetrics: [00:04:56] So this is always an interesting one for me, because it really depends on what you are trying to do with your marketing and what you're trying to improve. And so this number, this, what KPIs should you look at? It's something I get asked a ton, but it doesn't have a specific answer.

But what I like to do is look at what you trying to accomplish. And usually that means, look at a funnel, look at a marketing funnel that you have. Let's take it from a SaaS perspective where we have an ad. We have maybe an opt in for some sort of white paper or a guide or an email marketing course. We have a nurture sequence.

And then we start with the product or the trial after that. Um, some people skip right to the trial, but the idea is the same. You have a flow that people are going to go down . The main KPIs that you want to look for in a marketing funnel. Like this is how many people are coming in. So how many visitors are we getting?

How many opt-ins are we getting? So what's that conversion rate? What's the conversion rate to paid? And then at the end, how much money are we making? And everything else comes back and can be calculated from these just four things. And I think the key number that you're looking for is from the number of people who come in.

How much are they worth at the end of the funnel? And that's the number that you want to improve. So if I get a thousand people coming to my homepage and I made a thousand dollars, each person's worth a dollar. So then that, and that's my core KPI. I want to improve that number. And then as I look at different parts of that marketing funnel, that's when I'm going to be build out those secondary level KPIs.

Right. So, you know, you were talking about, okay, each department has different KPIs that they're watching and monitoring. Well, if we're looking at improving the marketing funnel in that nurture sequence, then there's going to be different KPIs that we're looking at. We're going to be looking at email, open rates and click through rates and retention rates and unsubscribes and stuff like that.

But that's not the core KPIs that we're looking at the top level. So when someone says, what KPIs am I looking for? It really matters. What's the goal? What are you trying to improve in your marketing funnel? And I said, I'd like to start at those, those top four or five. And then as we find a place that we're looking to optimize breakdown and break down into KPIs for that initiative.

Rob Walling: [00:07:24] That makes a lot of sense.  I would follow along with that, those KPIs of that funnel as well.  What do you see as the number one issue then when people start looking at marketing KPIs? Because I mean, I think our title is kind of saying, Hey, KPIs can be misleading or there can be counter indications or people maybe misuse them.

So what do we mean? What do you mean by that? You know, when you're, when you're saying that they can be misused or misunderstood.

Keith Perhac, Segmetrics: [00:07:48] I think that people think that KPIs are because it's in all the MBA books everyone's talking about, Oh, what are your KPIs? It's the number one thing that everyone thinks they need to focus on.

And I don't believe that. I believe that KPIs are kind of like what you were saying is a great indicator of how you're doing. Are you improving or are you not improving? Right. Are you going up? Are you going down? And that's all they are. They are a great measurement of where you are at any point in your marketing or in your founder and your SaaS journey.

However, they don't really tell you anything other than I'm doing better or worse than I was last week. Right. And that's the key problem that I think a lot of people have when they're starting to look at KPIs. So two key problems, one, they look at too many. They're like, they forget the key. They're just like performance indicators.

So like, okay, we need to know how many leads and how many people opened emails and how many clicks and how many people click this one little link and how many people are named Sally on Tuesdays. And like, these are all the numbers that we think we need to know. No. And as soon as you start measuring all of them, you have lost any key performance indicator.

And so the rule of thumb that I always go to is. You want five at the most, you want five key performance indicators for any department or any initiative. Right? So for example, if you're the CEO you want probably how much is revenue going up or down? 

Rob Walling: [00:09:22] CEO would be, but it might be net profit or it might be money in the bank.

Right. There's yeah.

Keith Perhac, Segmetrics: [00:09:27] And then you have your marketing department and your marketing department has other goals. And then your ad department has a different set of goals, probably around ad spend and cost per acquisition. But the CEO doesn't need to know, okay, what's our ad spend versus cost per acquisition right now. It's not his key focus. He has someone in charge to look at that.

And I think that that's the main problem that people come to, which is they just want. They think this is important. So it's a key performance indicator.

Rob Walling: [00:09:57] Okay. So too many. So that's the number one issue you see is just too many. Now you mentioned something to me about outliers and that how important outliers can be. Is it outlier, KPIs, or just outlier numbers are when trying to measure your marketing

Keith Perhac, Segmetrics: [00:10:10] Outliers in general. And this goes back to, I think what I've kind of feel is the second problem with KPIs is that people think that they can then improve their business by looking at those KPIs. Alright, so let's take the MRR, um, example and we're like, okay.

We had a huge dip in MRR. Okay. Well we know why, because a bunch of people canceled, but what did that tell us. That didn't tell us anything. We just know that we're doing worse. So people have this idea of, okay, I'm measuring the KPIs. Now I know what to do to improve the business. And that's not it at all.

It's like a speedometer. It just tells you kind of where you are. It doesn't tell you any problems that are going on. . So with outliers, what I always say is from that KPI, you want to find the breakdown of people who are doing really well at that KPI and people who are doing really poorly at that KPI.

So it's a standard one I think of is conversion rate because I'm a conversion rate optimization guy. I've been doing this for years. Let's say you have a marketing funnel, you have a conversion rate of 50% on that funnel. So 50% of people who come into that funnel convert. Awesome. You are doing great, but let's look at drilling it down one.

So you have two traffic sources. You have Google ads and you have Facebook ads. And let's say Google ads are converting at a hundred percent. And Facebook ads are converting at 0%. Well, suddenly that 50% great number now has a whole different spin on it. Right. We're seeing now that we're wasting all of this money on Facebook ads, because zero of them are converting.

This is not a hypothetical. This is something that we actually came to in SegMetrics, where we found that Facebook ads had a cost per acquisition of $800. And our Google ads had a cost per acquisition of $60. And it's like, okay, well we know where we're going to go with it.

And Facebook doesn't tell you, Oh, you this, because they only see their own data.

Right. They're not letting right. You're not able to see this data.

Rob Walling: [00:12:11] Right. That's fascinating. So I have a couple of questions. I think one, one comment I want to make is like, when I look at Amazon reviews, for example, like I can look at a review for a new spatula , a garlic press or something.

And I can see that it's like, A three and a half star, which isn't very good. Right. But I always dig into the reviews and look at what are the one star reviews. And there's often a bunch of one stars that are like, this was delivered late. And this thing, when I tried to bang a nail with it, it didn't drive it in.

And that broke, you know, and then you'll see there's a bunch of five star views in the, and the. Knowing that it's three and a half is actually not super helpful. It's helpful at, at a high level, but knowing the details of why there are ones and what people like and dislike about it, I think are important.

And that's kind of what you're saying here is mixing these numbers. I see this with, um, I've seen this with churn where everybody groups, Hey, my churn number, my monthly churn right now is 5%. And it's like, Oh, what's your churn for, you know, the first month? Uh, or the first 60 days or what's for the first 60 days of this particular cohort, you know, you can drill down those numbers are harder to get too.

I think today with dashboards, like ProfitWell, Baremetrics, ChartMogul, they're easier, but , it's hard to fully understand it and yeah, until you really get in and are able to drill down to that level, you're here often kind of looking at this aggregated data, which while that's sometimes the best we can do as marketers, it's kinda, it's kind of dangerous.

And that's what you're saying. Right.

Keith Perhac, Segmetrics: [00:13:35] Exactly exactly. And let's, let's take that Churn number. So I think a lot of people have realized recently that churn initial just the top level churn is not, not a valid metric because your churn in the first month is going to be so much higher than standard churn. And especially when you have a trial or something like that, that turn really doesn't. Count, I guess, but let's look at it one different way, which is okay. So now we know we're going to ignore churn for the first 30 days. We're only looking at what we'll call real churn. Well, now we want to know what's the churn of ads versus organic versus referrals, right? So now we've added a second level onto there, cause we still have that churn KPI, but we want to figure out, well, we have a 10% churn.

Who's churning? Why are they churning? So. Where do we start asking questions? And this is what I think is really missing. And what's important for marketing analytics tools, which is being able to ask these questions and in real time where I'm saying, okay, I see I have a turn of 10%, is it okay? Where are these people coming from outlier there?

So I look at my, my, uh, levels of churn for, out for a lead source and look at Google, uh, Facebook, organic, et cetera. And I say, okay, now they all look the same. Well, let's ask a different question. How about how far they get on the onboard and suddenly we see people who don't complete step two in the army boarding churn 20 times, as often as someone who completes up to step three.

So being able to ask these questions in real time and just see a chart, right. Just see a list of numbers of, okay, now everything looks normal or, Oh, look, this outlier is there and this is either really good. So we want to spend more to do that, or this is really bad. So we want to fix that. And that's, to me, the power of those outliers, which is, when everything's good and everything's copacetic. Awesome. You're doing great, but that's not what you're going to do to improve. If everything's good, you don't have anything to do. But somewhere in that good is something that's not performing well or that's performing great. And you want to do more of that.

Rob Walling: [00:15:44] Got it. That makes sense. And is that like, is there a major mistake that people make when they start to look at outliers?

Keith Perhac, Segmetrics: [00:15:51] I think they look at places. Where they don't have enough data yet, or they don't have focus. So one of the big ones that I see, especially with overall conversion rate optimization is so let's say we have that, that sales funnel let's do a product this time.

And people say, I'm just not getting any sales. We need to optimize the sales page, this checkout page, where they put in the credit card information we really need, because that's closest to the money. It makes sense, really want to optimize that. It's like, okay. Yeah. There's a lot of things. We can change on this.

A lot of things we can do to optimize this, how many people are coming to this page? Like five a week, like. Dude, you're never going to be able to opt out and ask because it's going to take, there's not enough data. So this is really the big issue that I see is people start way too far down the funnel.

When they don't have any data. So the kind of rule of thumb I use is to work backwards from because you want to be as close to the money as possible, but work backwards from that. So start at, okay. Check out. Do we have enough traffic there to, do you have enough people coming to that page to make a difference?

No, go to the next one. Sales page. How about there? No emails, no opt in page. Okay. We finally have enough people coming to the opt in page. But they're not converting. Okay. So let's fix that first and then lets fix the next one and then let's fix next one. And it's really, it is a funnel because the closest things you can get to the money are at the end.

But if you don't have anyone coming in the top, it doesn't matter.

Rob Walling: [00:17:24] That's interesting. Asia Orangio talks about starting your optimization at the bottom of the funnel, which makes sense, because if you get that working and as you work backwards and more people, you push through it, the more you're going to convert.

But your point is until you have enough people you can't really improve, how can you really even know how to improve, right. How to test, right.

Keith Perhac, Segmetrics: [00:17:46] Yeah. And I think there are of course, low hanging fruit that you can do on that checkout page.

You could make sure it has a button. There's obvious things you can do at the beginning. At the bottom of the funnel to make sure that it is following best practices, it works right. But when you're looking to improve right. Rules of thumb, but when you're really looking to improve, when you're not, when you're looking for that next thing mean you have to move up the phone and you have to have enough traffic to be able to test because I've worked with a number of clients who are like, Hey, we want to test this button on the sales page.

If it's red or it's blue, and this is a small change. And if you're not getting enough traffic, you're never going to get an answer to that question. But if you have something that has a 1000 or 2000 people coming to it, you're going to get that quick answer question. That question answered pretty quickly.

Rob Walling: [00:18:32] You know, I remember I, I tweeted something , it was probably five years ago and it was just a random thought I had  I'd said like being a founder is making hard decisions with incomplete information.

And then I said, I find most of the decisions I make it's like 80% is made on data. And 20% is like rules of thumb and gut feel or something like that. And someone came back, I don't know who it is, but he was like, no, a hundred percent data all the time. I make every decision with data.

And I remember being like, well then how do you build a landing page when you have no traffic? You don't have data yet, you know, it's like that, that's what I was trying to imply. That is. I use, um, my rules of thumb or other people's rules of thumb to do the first cut at pretty much everything. You know, if it's an email that I'm going to send out, if it's a landing page, if it's a, a product or an ad campaign, it's like, how do you decide the first 10 images to test test when you have no data?

You know, what 10 images should I test in the Facebook ad or whatever. And then once you have the data, of course, we're going to go with that, but there has to be something. So  gut feeling , I think comes from experience or  from, you know, an instinct or whatever.

But I think the nice part is rules of thumb. You can often hear on, on shows like this are conversations with founders who have been doing it long enough that they have their own mental rules of thumb, right. Of what, where the conversion should be.

Keith Perhac, Segmetrics: [00:19:44] And I think that that's also important when looking and it's, it's a dichotomy because marketing is part science as it's part database, and it's part art because you have to know.

What to do and how people think.  It's almost like a, like applied psychology in a way. you have to understand how people think like you can see the data all day.  Organic is converting really well, where Facebook is not, but that doesn't tell you what you need to do to fix it. And that's where the art and the psychology and the understanding of those, and that comes with experience.

And even the idea of like, okay, what should we look at to find those outliers? What are some things like, people aren't, they're coming on this webinar, but they're not converting. Well, Facebook ads are not converting because they weren't looking to purchase anything. They're not, it's like just thinking through that mental process of people in each of these modes, as they're coming through that funnel.

And that's how you start getting those ideas of, well, what about this? Oh, what about this? And that's, I think how you build those outliers, how you start to identify people who, who are performing and who aren't performing, and then figure out how you can best tailor your messaging towards them.

Rob Walling: [00:20:57] We have a question from the live audience so Rison Simon says, what is a good traffic number to start optimizing for conversion? So I think he's kind of asking like how much traffic do I need to start splitting thing?.

Keith Perhac, Segmetrics: [00:21:12] Right.  So I generally have around a 100 people. Going through something before I, um, before I start optimizing it. So traffic wise, I would actually say that so traffic is different than optimization. So what I would say is if you have a hundred people filling out the form on a page, you are ripe for optimizing that opt-in.

Yeah, 100 conversions at any step. Is what you should use to say, okay, I now have enough people to start testing this and going through this. Traffic wise, it's interesting because especially when you're split testing,  you have a calculation of how statistical significance of that test.

And what we have seen is that when you have small changes, the statistical significance is generally very small. So if you change the wording on a button, a lot of times the change is very small. The difference is going to be very small and it's going to take a long time for you to figure out, okay, did this really move the needle?

Where if you have a big change where it's like a completely different picture of a person or a person versus a robot, or like  H1 changes, right? So you're changing the key proposition of the page. These are tests that we see big differences in conversion rates, and you're able to get that statistical significance much quicker.

Yeah. Never knew multi rate, uh, testing when you're, when you only have a little bit of traffic. Cause that will take you years.

Rob Walling: [00:22:44] Never. Never. Yeah. Yeah. And that's, um, that's a big part of it, right? Is it's that experience level of, I know that if I change the headline or the design, like everything on it, that I'm going to have a dramatic difference.

And I know that if I changed some wording in a body paragraph or I change one icon or I changed the button color from blue to red, the odds of me getting that big of a change are pretty low. So you need more traffic, you know, at that to make these tiny tweaks. I think that's a mistake. A mistake I've seen people make is not changing enough or not, not making drastic enough changes.

And then you just let the test run for a month or two. And then it's like, Oh, these are basically the same thing.

Keith Perhac, Segmetrics: [00:23:27] Yeah. And if you have low traffic generally focus on, so there is no more fold anymore. Anyone who says, Oh, it has to be above the fold. Doesn't exist anymore to a degree, but focus on  that hero image that above the fold, what is your H1?

What is the iconography that you're using? When people first visit the page, that's going to be your biggest win because that's the first experience that people are going to have with that page. .

Rob Walling: [00:23:52] Very cool, sir.   We are almost at time , but do you have any other thoughts before we close?

Keith Perhac, Segmetrics: [00:23:59] I think that I think, especially in the developer community and in the founder community, I think there's a lot of people who think, Oh, marketing, I don't get it.

Or, Oh, it's a dirty word kind of thing. I I'm a developer by trade. Um, I kind of fell into marketing and. The biggest successes I've had with marketing is thinking about marketing the same way I think about development. Where it's not a miasma of things I'm going to do. I break it down, just like we were talking about with those funnels of, we have a start condition, we have an end condition and we have a number of steps, processes that are going to go through everything.

And we just put a number on each one. How is this one doing? What is the value of people who have done this step? What is the dollar value of people who've done this step? And just writing out. I have flow charts for every single marketing campaign we have and we put a dollar value to every single thing, and that lets us easily see where our holes are.

And we can really look at it in a more scientific and structured way than I think a lot of people who are not. In that marketing world, uh, realize that you can do.

Rob Walling: [00:25:08] Yep. I agree. I'm also a developer engineer from, most of my life. And that's how I view marketing too.

I've always viewed it as a technical marketer, a left brain marketer where the numbers in the funnels made sense to my brain. Um, I've heard marketing described as math and human psychology.   the math was easy. The math is very natural. And I think if you're a developer lean into the math and learn enough of the human psychology, you mean when, when you say human psychology, it's how people make buying decisions, the stages they go through, , what copywriting is and what sales copy actually looks like.  These are the things I had to learn and I read all the, I mean, this was 2005 to 2011. When I was teaching myself this, it was a, there was no startup writing on this.

There was no CopyHackers at the time. It was all the info marketer guys. It was that. Dan Kennedy and that whole crew, which I don't love a lot of the stuff they do from personal perspective. But, but cause they're just too hard driving sales and it's money, money, money. And I feel like some of them like trick people into buying stuff.

Um, but they know copywriting and they know persuasion. And that was a big thing that I think is kind of the other side of it. Totally. And then you can learn from them and apply it in a way that feels comfortable to you. And that's all it is. This is not, it's not so complicated. It's not black magic.  the first copy I wrote was crap. Didn't convert very well. And the 10th, , landing page I built was really good, you know, and it's just, it's just learning like the first hello world program you write sucks. And then  that 20th  code you ship is much better.

Sir, Thank you so much for joining me on MicroConf On Air today.

Yeah. Your Twitter handle is,  I feel like I'm going to pronounce it wrong cause it's Japanese. Isn't it. is @harisenbon79.

Keith Perhac, Segmetrics: [00:26:38] Very close.@harisenbon79.

Rob Walling: [00:26:41] Got it. Okay. @harisenbon79 You used to live in Japan for those who don't know.

So thanks again for coming on the show, man.

Keith Perhac, Segmetrics: [00:26:49] Thanks so much, Rob. I had a great time.

Rob Walling: [00:26:52] Absolutely.

All right. And a couple of housekeeping items as we check out. Thank you so much to Hey and Stripe. There are a flagship partners with MicroConf for the year and they make everything. We do just a little easier and we always appreciate being partners up with them.

MicroConf masterminds are happening. Round two of our matching go to MicroConfmasterminds.com. I believe it's plural, or you can just head to, MicroConf not common and search around for masterminds. If you haven't already had to youtube.com/microphone, it's scribe, and we'll see you next Wednesday. Same time, same place.

MicroConfOnAir.com. Have a good week.