November 28, 2018
Ep. #7, Observability at Asana and Honeycomb
In episode 7 of O11ycast, Charity is joined by fellow Honeycomb team member Michael Wilde along with Asana's heads of engineering and tech, ...
Let's chat about conversion optimization. Before diving in, I just want to give you a little bit of background on myself and why I'm here. As Dana mentioned, I'm the founder of Growth Pilots, which is a paid-acquisition agency. We focus on helping high-growth companies scale their pay channels.
Paid acquisition has a lot to do with conversion optimization, obviously, because the two are very interrelated. You drive traffic into the top of the funnel, and you need to be optimizing it all the way down from channel all the way through to conversion.
Over the past five years I've helped quite a few companies with various aspects of growth. There are some of the companies logos up there on the screen: WordPress, Kissmetrics, as Dana mentioned, Hiten, who referred me here, ZenPayroll, Intuit, etc.
My big thesis is I really see there's this big opportunity in helping companies figure this stuff out. It turns out that companies aren't very good at figuring this stuff out on their own because you don't really know how to optimize a conversion funnel unless you've done it several times.
Most entrepreneurs have only done this a handful of times at most. When you start getting large data sets, it starts to make a lot of sense to help companies do this, and that's me.
Why optimize in the first place? I think that lays the groundwork for this whole presentation here. Hopefully, I don't need to explain why you should be optimizing. But it's actually not always that clear.
Optimizing is one of those things that a lot of companies and entrepreneurs put on the back burner. It's like, "We have this landing page, and yeah, we could optimize it. We could do better. But there are higher priorities."
Sometimes that is true, but at the same time, optimizing your conversion funnel is what ultimately leads to you getting more efficient in your funnel and getting more customers. So that's really the impetus for doing conversion optimization.
You want more customers, right? But, at the same time, on the opposite end of the spectrum, there are companies who over-optimize. Over-optimizing is to try to optimize every little thing, every little color of buttons.
The thing about optimization is that there's this quick-diminishing returns. So, the first five tests you do are going to be much more effective than the next five that you do, and so on and so forth.
As you start testing, over time you're going to see that things tail off pretty quickly. Finding that balance of where you should optimize and how often you should do it, it's on a case-by-case basis. But it's something to be cognizant of.
Then it comes down to, when you're doing conversion optimization, what should you focus on? What should you optimize? And it's actually really simple.
There are only two things that you need to optimize: there's how people find you, and then there's what they experience once they find you.
That's pretty abstract, I realize, so now let's just visualize it a little bit. This is the conversion funnel. You have channels, and this is the "how people find you" component, and then you have "what people experience once they find you," which is a landing page, the onboarding funnel.
The end goal here is to maximize all of that and optimize it to get to a successful product experience, which is a conversion. One of the big misconceptions here a lot of companies focus on is optimizing everything post-channel, meaning, "Okay, let's just look at the landing page. Let's just look at the onboarding flow."
But actually what you find is the biggest lever of conversion-rate swing is at the channel level. What I mean by that is, by testing different channels, you'll notice that there's much more to be gained from a conversion-rate-lift perspective than there is by testing your landing page or your onboarding funnel.
The goal here is really to keep these things consistent. It's kind of an interesting way to think about it, but you should really think about your conversion funnel as a story, right?
You want consistency and relevancy from the time somebody first hears about your product or sees it on some third-party channel. Whether that's SEO, whether that's SEM, whether that's a blog post you're writing, whether that's a referral that somebody, an existing, happy customer, has said to someone else.
Whatever that first perception of your company is, you want that to follow all the way through. That's how you maximize conversion rate, and so you want that to be consistent from end to end.
The best way to do this is to start with your product experience and work backwards. You know your product. You have more control over your product than anything else, and so from there you can extrapolate and back things out and build the conversion funnel backwards.
This is the way that I'd actually think about the conversion funnel when you're thinking about optimizing. You want to start with your product experience. Once you have the product experience, you can then bake that in, pull everything back and say, "Here's what my product does. What would I need in the onboarding experience for somebody to believe this story?"
To believe that, they should go from the onboarding experience to successful product experience, which is a conversion. And, likewise, to tie the next two things together, you want to go from landing page to onboarding.
What does somebody need to see on your landing page to make them want to continue through the funnel and hit the onboarding process? These things all tie together, and you can't ignore the channel level, either. As I mentioned before, this stuff goes all the way down the stream.What somebody sees on an external channel, you want them to have a good experience and, really, there should be relevancy there from whatever they see on that channel to what's on the landing page.
If your channel says something, let's say you write a blog post or let's say somebody posted about your product in a developer forum. Whatever that person says, that should reflect whatever's happening on the landing page. Because what's going to happen is you're going to get drop-off if there's not that relevancy factor there.
Now, just taking a step back and talking about channels in general is the very top of the funnel.
You want your product to have this relevant message across all of your channels. And that might be different for each individual channel.
When you're doing email marketing, you have a list of prospects that you're emailing, or leads. That message that you're sending, you might want to have a very different experience, and even a different landing page, for them versus SEO, which is more organic.
You have less control here. The channels that work best are the ones where the experience is the most consistent from the time that somebody first sees your product all the way to the landing page, to the onboarding flow, to the successful product experience, the conversion.
What that actually ends up looking like is the channels that you have the most control over are the ones that are the most effective because you can influence what somebody sees on a given channel.
Some of these channels are very dynamic. You can't actually control the messaging there, and so your ability to influence what somebody perceives about your company is going to be, well you're just not going to be able to do it very effectively as opposed to a channel like SEM, paid acquisition. You can effectively tailor that message to be whatever you want.
You can have a custom landing page, and so you can tell the story that you want. That's really the key when it comes here. I just listed some channels here, just in case, to chat about these a little bit. Hopefully, these are all clear.
SEOs: Search Engine Optimization. Content, which would be more on the social-media side of things and creating content on your blog. Email: a really effective acquisition channel for SaaS Companies tends to be lead generation or getting these lists and sending these emails out to companies who could potentially be interested and putting them into the sales funnel. Paid acquisition, of course, PR and word of mouth, which is a referral.
On that note, word of mouth is an interesting one because you don't have very much control over what's said about your product by somebody else.
If somebody's talking about your product, more than likely they get it, and if they're going to be an advocate of it, there's probably going to be a pretty high correlation between what they say and what someone's going to experience.
That's one of those outlier channels that doesn't really follow the norm, and as a result, it tends to be a very powerful channel.
Moving from channel to landing page. There are some core elements to a really strong landing page, and we're going to walk through some examples of this. Landing pages that have these elements tend to be the landing pages that convert best.
The first thing you want is a strong headline that communicates the product value.
This is your value proposition. Again, this comes down to relevancy. Why is somebody on your page in the first place? Tell them the benefit that they're going to be getting from this.
There's credibility and third party validation. Are other people using this? Are companies or people that your prospects know using this? This comes in the way of customer logos, customer testimonials, and so you tend to see a lot of that on landing pages.
A product preview, in keeping with this whole consistent story theme, you want people to be able to see what are they going to experience at the end of the funnel. What does that product look like? Why should they continue to believe your story and buy into this all the way through to the end?
And then, lastly, a strong call to action. Tell people what you want them to do. I've seen a lot of really great landing pages, and then they fail to ask for that selling point. You know, "Sign up now. Free trial. Buy now."
A lot of people miss the mark on that, and they'll hide a button up in the top right corner that should be front and center once you have all of these elements working together.
From there, moving on to the onboarding flow. In keeping with this whole relevancy theme, you want all of the steps in your onboarding flow to align with the product expectation. Drop some hints along the way, maybe some screenshots about what you should expect once you hit that successful product experience.
It's really about minimizing friction, right? Don't put unnecessary steps. If you don't need a field, don't add it. That's just another reason for somebody to drop off and not complete the rest of your funnel. In general, fewer steps is typically better.
There are some exceptions to this. When you have a product that really requires a lot of buy-in, and there's a lot of information that's needed, simple can actually, ironically, be worse, because it doesn't feel like a real experience. But, in general, fewer steps is typically better.
Then this is another big one that a lot of companies just miss the mark on: guidance along the way. You should really think about your onboarding flow from a fresh-pair-of eyes perspective. If somebody landed on the onboarding flow, and they didn't know what to do, would they just leave? What kind of guidance are you giving them, whether that's tutorials or tool tips, or even videos?
We're going to look at some of this in some of these examples, but, again, minimizing friction is the biggest piece of optimizing your onboarding flow. Just to do a walkthrough, is everyone here familiar with Fastly? Fastly is this new content-delivery platform, and they're a very hot company right now.
I figured this would be a relevant company to do a walkthrough from a conversion-flow standpoint. Here's their landing page. As you can see, we're defining content delivery. That's really their core value prompt. That's the thing they want you to take away from this. Then there's some supporting text underneath it, and it explains a little bit more in depth what they do.
On the right you can see this really cool meter, and this is just a screenshot here, but that's actually a moving meter. This is that expectation of what the product does. This shows how fast and how optimized their platform is. It has all of those components.
You can see this strong call to action, "Try Fastly Now," it stands out. Your eyes definitely go there, and then, at the bottom, you can see the list of companies that are using them. They really hit the mark on all of those elements that I talked about when it comes to a winning landing page.
This is the bottom part of their landing page, providing a little bit more clarity, a little bit more product expectation, and then on the bottom, another customer testimonial. Great. So, this is step one of their onboarding flow, right? Once you clicked the "Try Fastly Now" button, you end up here, and you can see they're giving you this really clear guidance of what you're going to get.
You talk about expectation, they're literally telling you a bullet-point list of what you're going to get, and they show you a screenshot. They definitely hit the mark here.
One other really interesting thing is they continue to have the "Try Fastly Now," call-to-action button. If I were consulting fastly, I would tell them to remove that, because somebody's going to click that, and once they do, it's just going to go in an infinite loop there.
Once you get through that step they send you an email, and you go through this next step of the funnel here, and again they really outline what you should expect to see: step one, activate your account; step two, configure your site; step three, learn more. Just providing the clarity and expectation. Just guiding the user along.
Very strange to see the call to action being there again, because once you click that, you start all over. But outside of that, this is a really good onboarding flow.
Once you confirm your account, this is the page that they drop you on, and this is kind of the quick start to get things up and running again. Very simple, three fields and then configure. A lot of guidance on the right hand side. If you got lost you'd know exactly what to do here, and so they're doing this well.
This has definitely been optimized, and from there, a quick start. A lot of instruction here. Guide to tutorials and anything you need. If you got stuck on this you would know exactly what to do, and that makes for a good onboarding flow.
Then they also send you this really cool email, so just in case all of that isn't enough, they send you an email that outlines all of the steps that you need to do in order to get started. They're hitting you from all angles, and this really allows you to build this really effective, multi-touch funnel.
The second company that I want to walk through is actually a client of ours, and I don't know if any of you have heard of them. They're called Mode Analytics, and they have this really cool data platform that's built around sharing SQL queries with anyone in an organization and empowering non-technical people within an organization to run with the SQL query, and run it whenever they need it, and even edit it a little bit when they otherwise wouldn't have been able to.
They just recently did a rebranding here, and this their new optimized experience. Their first value prop that you see: who, what, when, how and then answer why. Your data has a story. Mode helps you tell it. That's the value prop. Pretty strong value proposition there, and then you have the call to action there.
This is a scrolldown of the same page. Here's the product preview, again like we talked about, showing what you're going to expect once you get through that point. What is that product going to actually have once you unbox it and you get to the end of the conversion funnel?
Then, at the end here, there's that customer validation piece again, two key, marked customers who are using this. And then, if you notice at the bottom, there's always this little chat icon making sure if anyone has anyone questions, or gets lost along the way, they can ask somebody.
Step two: as soon as you click on the call to action, it brings you here, and again, you can see they really keep you engaged here. A simple field, email address, or you can sign in with Google, and they have the customer testimonials on the right-hand side. That further validation of keeping people in the story and giving them a reason to continue. Great.
From there, again, very simple, four fields keeping the customer validation on the right-hand side, making sure that people are staying engaged, giving them that reason to continue. Then you'll notice you continue to see that chat dialogue there. So you have any questions, boom.
And then, lastly, optional fields are really interesting. This is one of those things that you can definitely test and optimize. An invite flow is very important for them. They want you to invite other people from your organization, and not only do they want you to do that by giving you the fields, they also tell you why that's important on the right-hand side.
Giving that context, "Why should I believe this? Why should I do this?" They answer that really well on the right-hand side with plenty of explanation.
Here is connecting your database. That's the last piece to doing this, and again, really strong tutorials and description on the right-hand side of why you should do that. Finally, entering your actual credentials. In this case I just chose Redshift for example, and then, again, there's that nice tutorial on the right-hand side that tells you what you need to do in order to complete the onboarding.
You should never get lost along the way, and if you did. you'd have this opportunity to chat with somebody.
In totality, once you're done, they drop you on this really cool tutorial screen where you can see all the different things that you can do and all of the different reports you can run. Things that you can learn by using Mode, which, again, this might not seem like funnel or conversion optimization, but it absolutely is. That's the next thing I want to chat about here.
Choosing what you're going to optimize is often one of the most difficult things. Sometimes you might think you want to optimize for signup, right? Or you might think that you want to optimize for some kind of key action. Maybe, in Mode's case, it's getting somebody to invite other people.
The problem is you don't really know which of these that you want to optimize for. In an ideal world, you're obviously choosing high-value customer. If you could, that's the one thing you'd optimize for, more customers and the best customers. But the feedback loop on that is so slow that if you were to optimize for that, it would take you two years to end up actually doing any optimization.
That's really the tradeoff that you have look at. When you're choosing what you want to optimize for, there's this feedback and signal tradeoff. There are events where there's this instantaneous feedback. A signup's a good example of that, but the predictor of a signup on the quality of a customer or the quality of sign up is very low.
Maybe you don't want to optimize for that, and then if you move on to this next step here, some kind of key action. That's post-signup. Somebody's engaged. Again, maybe in Mode's case that's inviting a customer. That happens faster than a signup, but maybe it's not the best predictor of whether somebody's going to become a high-value customer or not.
As you go down the list here you can see getting to a high-value customer takes a long time. It's the best signal because if some channel, some landing page, brings these high-value customers, you want to do more of that. It just takes a really long time to get to that point.
You have to find this balance, and this is different for every business. You have to find this balance of the action that you want to optimize for, that's the most valuable, that you can get, where the feedback loop is short enough for you to be able to have actionable insights and actually optimize against.
All of this optimization talk is great, but you can't optimize unless you're measuring this stuff. So when you're thinking about measurement, you want to be able to measure what happened. That's that key event.
Is that signup? Is that a key action? Is that a paying customer? Is that a customer invite? That's really the crux of what you want to be measuring.
Arguably the most important thing is measuring the "where" component. Where did that traffic come from? Because that's where your opportunities are for optimization. If you treat all channels and all traffic the same, you're not going to be able to do very much granular optimization, because the channels all perform so differently.
Traffic performs so differently that you want to be able to segment this stuff and see where it's coming from in order to really unlock all those inefficiencies in the funnel and optimize them.
Of course, the goal of all of this is just to really understand why an event happened and then take that data and optimize it, adjusting to improve that. That's all optimization really is, right?
Measuring, adjusting, measuring, and then rinse and repeat.
Then there are more robust tools which might be overkill depending on what your product is. It just really depends, and these are more database or SQL-driven. This would be like Mode Analytics, the Periscopes of the world, Keen IO, etc.
In order to feed data to these measurement tools, there's a concept of UTM parameters. So, how many of you are familiar with UTM parameters?
UTM parameters basically let you plug different parameters into a URL in order to tell some system components about that traffic.
You can tell where and what that channel is, whether that came from Facebook, whether that came from your blog, whether that came from email. You can even break it out by campaign. Is this an email newsletter sent on July 15th? The more granular you can be here the better. Because what's going to happen is you're going to start gathering all these events, based on all of these different UTMs, and then once you start segmenting and looking at all of this stuff, you'll start to see what's most effective, not only from a channel standpoint, but even from a campaign.
You'll start looking at the different dates things were sent. Maybe you used a different message in your email newsletter this time around. Really adhering to UTM Parameters is something you absolutely have to do if you're going to do proper measurement and optimization.
In talking about UTMs and measurement, there's this concept of attribution.
Attribution is how you credit a customer or a conversion, or whatever that key action is, to the various channels that that conversion touched.
This might sound pretty simple in practice, but as you can imagine, when you start getting traffic from a bunch of different sources, it's kind of confusing about who should get credit for that. If somebody searches for you organically, and they hit your site from an SEO, organic Google listing, boom. They hit your site.
They don't convert, they bounce, but then you write a blog post, and they end up clicking through to that. What should ultimately get credit for the signup or the conversion that happened? Should it be the blog post? Should it be the organic ranking? That's where things start to get really tricky, and it's really messy, to be honest. It's not easy.
There are different philosophies or methodologies here. There's what they call "first-click attribution," which means to assign all of the credit of a conversion back to the first source or channel that it touched. In the case that we chatted about, that would be that organic traffic would get 100% of the credit.
On the other end of the spectrum, there's "last click," which would be the final click that took place. So your blog post would have gotten all of the credit in that case. And then there are all kinds of in-between stuff.
There's uniform, which, maybe there are five channels that touch a given conversion, each of those would be weighted equally. And then there's parabolic, which is where the first touch point gets a large majority of the traffic, and then it equals out from there, and the last touch gets an equal amount as the first touch.
There's a lot going on here, and on top of it, there are cross-device issues, right? If somebody clicks from a mobile device, and then they go to a desktop, you're not going to be able to effectively track all of the attribution tied to that.
The UTMs break off. Cookies aren't stored from device to device. Now, that's getting better. Facebook's ad system is actually really good at that. Google is not so good at it. Then there are just cookie deletion issues in general. If you have a really long conversion cycle, if it takes three months from the time somebody signs up to convert to a customer, more than likely you're going to see a lot of attribution issues. Because people delete their cookies or they change computers or browsers, etc. So, a lot of issues there.
My advice here is really to just keep it simple initially, choose what makes the most sense and don't spend too much time over-thinking this at this point. This is definitely a later problem. Once you guys all start spending millions of dollars on paid acquisition, which hopefully you do, then you can start worrying about this and really dissect how it makes sense to treat conversions from an attribution stand point.
Now, I just want to chat a little it about SEM. That's really my expertise, and while I don't think that a lot of companies here are prime candidates for scalable SEM, there are a lot of really interesting clever things you can do to leverage it, even as you get up to that point.
Is everyone familiar with what SEM or AdWords is here? Show of hands possibly. Okay, cool. Yeah, mostly everybody, right? It's an auction-based ad system where you bid on keywords. So, when people type in relevant keywords, you can show an ad against that. A lot of control there.
You can choose what kind of traffic you get. You can choose the ads that you show to them. If we go back to "control over channel," that's a great channel for that. Because you can really influence what somebody perceives about your product and then bring them to a landing page that's relevant.
SEM can be a really big and meaningful channel. For a number of our clients it's their number one channel, meaning this drives the majority of their customers. It definitely can be something very big and scalable. But it doesn't work for all companies, and you definitely have to keep that in mind.
You know, I see a lot of companies continuing to chase up the tree and try to make SEM work, but it just doesn't work for all businesses. Whether that's because your unit economics aren't great, and you can't justify the competitive prices that the bids are set for these keywords so your LTBs don't back out. Or your market size just isn't that large from a search standpoint. People aren't searching for what you offer. It's important to keep that stuff in mind.
But yeah, again, that's just a quick overview of SEM. I think the crux of what I want to talk about is really the tactics that you can use at this stage, kind of at the early stage, to leverage SEM to add value to what you're doing.
I think there's this handful of things that are really meaningful that you can do with SEM at an early stage, and one of those is testing messaging. If you're trying to figure out what your value prop is or you want to see what converts best, maybe you're going to put that headline on your landing page.
Actually testing that within AdWords is great because you can pay for clicks, pay for impressions, and you can see what the relative clickthrough rate is on these. Do an AB test, and you can see, "OK, this messaging performs better than this messaging." So, maybe you draw some hints from that, and you apply that for your landing page and the rest of your funnel.
And then targeted traffic for experiments, maybe you're launching a new feature of your product, and you want to see how receptive people are to it. You could launch a campaign around and buy keywords that are related to that feature and see how receptive people are to it.
Are people signing up? Are they converting? It can be really helpful for these smoke tests or early validation to see whether or not that's something you should roll into your overall product.
This is an interesting one, estimating customer-acquisition cost. A lot of companies don't know what it's going to take to acquire a customer. They don't know what their conversion rate is. They don't know what they're ultimately going to have to pay for it.
So, as you're fundraising, this kind of stuff can actually be really helpful. Investors love to see this kind of stuff, what it actually costs to acquire a customer should you want to pay for customers. It's just a great way to see what your CPA, cost per acquisition, is.
Hijacking competitors' traffic: this is a little bit more gray-area, morally speaking. You'd be surprised how many companies do this. Even in the developer-tools community, start searching for different companies. You'll see, when you search for a company, that there are other related companies trying to steal or conquest their traffic.
A good example of this is actually Kissmetrics and Mixpanel. When we were working with Kissmetrics, you notice Mixpanel was actually trying to bid more than Kissmetrics was for their own branded terms, and that's very expensive in the AdWords auction.
AdWords rewards relevancy, and so when you're bidding on a competitor's term, it's very expensive. But, at the same time, those customers are probably very valuable. So you might be willing to overpay.
It's a little bit more advanced, but retargeting traffic on the search network, what this allows you to do is, when people visit your site, everyone here is familiar with retargeting I presume, but that's basically when somebody visits your site, you can show them an ad around the rest of the Internet and try to get them to come back and convert. It's very effective. It's a very cheap way to acquire customers and re-engage those people that have visited.
One of the lesser-known forms of retargeting is actually search retargeting, and why this is really cool is, when somebody visits your site, and then they leave and they don't convert, and they're still searching Google for, let's say, adjacent solutions, it can be really effective to show them an ad again. Because they didn't convert. They're still looking.
You could say, "Didn't find what you like? Come back, we have new features, etc." Very few people take advantage of this, but it's one of the cheapest forms of traffic on Google's search network.
That wraps up my presentation. I'm happy to answer any questions around any of this stuff. You know, I'm here to help you guys ultimately, so ask any questions around the presentation or SEM, anything in general, I'm happy to chat about anything related to marketing. I'll just go ahead and hand it over to you guys.
Regular retargeting is not based on what somebody searches. So, regular retargeting is, somebody comes to your site, and then you can show them an ad on a display network. You'd use an actual banner ad for that. Whether that's on Facebook, whether that's on some kind of display network or ad roll, or whatever you use. It's not tied to what they search, necessarily.
Retargeting for search actually is a Google search ad unit. If somebody visits your site, and let's say they bounce, and then let's say they're searching for a key word that's not your brand but related to some other key word, you can show your ad there, and it's going to be cheaper because, in Google's eyes, it's more relevant.
LinkedIn's great because it's very targeted. Particularly when you're targeting somebody based on their title or their industry, you can't get more laser-targeted than that. It's a smaller channel. It doesn't scale very well because there are just not that many people on LinkedIn actively looking at the stuff. And, again, that's a relative question.
Scale can mean something different to different companies. But yeah, LinkedIn's great for that. You know, Twitter, not to knock Twitter or anything, but we've tried Twitter for a number of companies, and we haven't seen high purchase intent there or even conversion intent.
While engagement's great, I think if you're doing Twitter, you really have to set your expectations about what you're going to get out of it, and again, this stuff is different for every company, and you have to test this stuff out. But, in general, I think the intent on Twitter isn't as ripe as it is on other channels, even like LinkedIn.
That's one of those things that is really best to think about it in terms of themes. We've managed accounts where we have thousands of different landing pages based on the keyword that we're running in Google, and if you can do that in a templatized format, great. That works well. But to do it in a low-hanging-fruit, I'd think about it in terms of themes.
What are the core themes or value props that you want to present there? You can get that down to a handful, and those can still converge into your core product experience. You don't have to have a different landing page to call a channel, for example. Like, "Oh, hey. Saw you came from LinkedIn."
It should be more based around those value props versus getting really finite and granular about a key word or a if it's a individual blog post they came from.
That's how I'd think about it. I'd bucket it in terms of themes.
Where we've seen Twitter work really well is mobile app installs. But when it comes to the lead card or driving someone to your site for the purpose of conversion, unfortunately we have not seen great results there. It's been pretty hit or miss.
Twitter's not a super expensive channel, so if you do a test budget of maybe 5x whatever you're willing to pay for a customer, you should be able to get a sense of whether or not the channel has any legs at all.
She asked if buying traffic to an article that's about you, by a third party, is effective versus sending people directly to your landing page, and what does that drop off look like? Interestingly enough, we've seen much higher clickthrough rates from doing that, when you send somebody.
You have this third-party source. Depending on what channel you're on, like on Facebook, for example, let's say it was a TechCrunch article. You can actually target people who are interested in TechCrunch, and you're going to see a much higher clickthrough rate, because there's that relevancy factor there.
From a conversion-rate standpoint, I've seen both sides of the table. I don't think there's a clear consensus on whether that converts better, but that establishment of credibility, I would say you'll probably see a higher conversion rate from traffic going from that testimonial or that article to your landing page, or higher conversion rates, than you will from somebody clicking directly on the ad a lot of times.
But there's still that in-between, people who are clicking on the article and then dropping off. So, there's still that step in the funnel where people are dropping off from the time they clicked the ad. They get to the article, and then they don't click through to your site, versus those people who actually hit the article and click on your site.
That conversion rate will likely be higher than the direct-from-Facebook ad, a lot of times.
The data you'd have to look at for that would really be like looking at like the LTV of somebody who came from one of those channels. I will say the conversion rates tend to be very high for people looking for competitors, and it might not be very cost-effective, because you're spending a lot to get there.
But the conversion rate tends to be high, and there's a clear correlation between the conversion rate and how similar your product is, compared to what they're searching for, as you can imagine. Because what you find is people will bid on adjacent competitors, even if they have a solution that's very different.
They'll still bid on them because that traffic's relevant to them. That conversion rate's going to be much lower, but, for example, like a Mixpanel and a Kissmetrics, those two are obviously very neck-and-neck when it comes to certain features. So, the conversion rate there is going to be higher.
That's a good question. He asked about companies, that have more of a demo call to action versus a straight Web funnel, and that's not that atypical at all. You have a landing page? You have a lead form, and then do you email them, or do you set up a call with them? What's the next step from there?
[For email] your funnel then becomes, I wouldn't want to say broken, but it becomes a little bit disjointed. Because your funnel now includes that email. That's then something else you can optimize. What are you including in that email to get somebody to do that next conversion action or that next step in the funnel, rather?
That's the way I'd think about it. The landing page, in that case, is step one. Your email nurture is going to be the onboarding funnel, if you will. That's really what you should focus on optimizing, and there are a lot of great tools for testing that. SendGrid, MailChimp, etc., for sending more of that transactional email once you have permission to do so.
The question was, "What do you do in the case where you can't pass a UTM parameter and get full attribution on this stuff?" You don't know where that traffic's coming from, and that problem hasn't really been solved.
Google Analytics tries to do some of that. All you can really do at that point is tie things back to a refer level and look at it from that. There's, unfortunately, not much you can do, though. That's why the more control you have, the better.
If you can leverage UTMs, any channels that you can use a UTM on, and, even if it's a channel that might be organic, I've seen a lot of interesting uses of this too. Maybe even a press article. A lot of times you can give a journalist the UTM parameters so you'll know the traffic, where that's coming from.
You'll know the article it's coming from. You'll know the date, the campaign, etc., and so you can track the effectiveness of that from there. When it comes to those dark channels, as you called them, it's not easy. You know, there's not a clear way to do that outside of just looking at the referred data.