Overfuel Cofounder and President Alex Griffis spent some time discussing website performance, search, and digital retailing in a recent sitdown.

What are your thoughts on AI today?

This has been a sort of a big topic of debate. I recently just did an article with Alex Melan with Smart Sites about the five predictions for this year and I think there's no denying that AI has been all the rage and that all these companies want to get in the race to get a product to market. But I think that inherently proposes some issues, right? If you're moving fast, you're not really asking the question, like, what's the problem we're trying to solve? And as a result of that, I don't know if you saw the article about the Chevy dealer using the AI chat and somebody convinced that AI to sell them a Tahoe for a dollar.

I think that was like a clear example of sort of a rushed product. I think there's a lot more behind-the-scenes operational stuff that AI could do. That could be simple things like predictive maintenance, proactively reaching out to customers, and setting appointments. There's a lot that could be done on the sales enablement side.

I think that's less customer-facing until the customer-facing products get a little more mature and catch up. I really think there's a lot that could go on behind the scenes that would that would benefit dealers, especially around inventory management. If we started, you know, crunching numbers and sales data and more competitively pricing cars. There's a lot AI can do behind the scenes.

How can AI improve the customer experience?

There's a big portion of the customer experience that could be dramatically improved. I think the focus on chat is a bit of a mistake because with live chat, the intent, or at least the expectation on the consumer side, is that it's going to be instant and intelligent and actually returning dialog or answers that are relevant to what they're asking about. But if you think about text messaging the expectation is a lot more casual. 

I think you could have the conversational APIs perform much better in a text messaging environment if you're sitting on the data in the analytics like we are with our products. We know every single step a person took on a website. We know every car they looked at, every button they pushed, and I think if you approach that from a text messaging standpoint, you can contextually offer them things based on what they've done. It kind of takes some of the guessing out. You have more of the history, you have more of an understanding of who the customer is.

A lot of the live chat products right now are just kind of sitting on top of the website, They don't necessarily see all the other interactions that are happening.

How can dealers improve the online path to purchase?

Well, there are quite a few pieces of that. I think that speed is going to become increasingly important. Over 80% of traffic that we see is originating on a mobile device. I think the entire experience needs to be built mobile first and really tailored to that form factor. What we're seeing today, if you pull up the dealership website on your phone, you have all these third-party widgets that are all competing for your attention and they're popping up left and right. They're slowing the website down. All of that has downstream implications. It lowers the conversion rate and lowers the overall quality of the website, So it hurts them from a search engine standpoint. It leads to higher pay-per-click costs. We look at speed and form factors, kind of the two most important things, you know, parts of the customer experience.

We're not anti third-party widget. A lot of OEMs have various requirements and plugins and tools that they use. But I think we're looking at it from how do we get a customer onto the site? It feels instant, it feels like a mobile, you know, native mobile application on their phone. How do we make the calls to action super clear and concise?

When we look at our data, it shows that eight out of ten people don't even scroll 25% down a page. How do you minimize all the distractions and bells and whistles and really just get somebody focused into the funnel? Hit search and type in exactly what you want, or click SUV, or click a sunroof, or any number of the features that you want. You just have to put it right in front of them and make it super simple. By doing that, we have found that most people don't scroll. Most people don't like to go to the second page of the search results. Typically, you've got to present the car to them pretty quickly at the top of the page to keep them engaged.

What CTAs work on a VDP?

That's why we lead with typically something like "lock in the best price" or "unlock manager special". Those tend to perform really well because people like a good deal. They like a coupon. The most leads that we collect through our websites is actually through the payment calculator. 

When you think about the sites you go to today, it might ask you to know your location and you're like, Why would I do that? Why do you need my location? Then it's going to ask you to fill out an application to get preapproved. Well, why would I do that? I haven't even found a car I like. Our line of thinking is to get them into the bids and present the most obvious options. Show them enough to submit a payment, let them customize that payment, and hopefully get a lead out of it. 

Then, once they have an understanding of their buying power, we hit them with "get a free credit score check". Our partners at 700 Credit integrated with us via API, so we can pull on our credit score and give them a pretty realistic number of what their payment is going to look like. So we lead with the payments, we lead with an action that instills some sense of urgency, whether that's unlocking some kind of a coupon, or holding the car for some period of time. We move some ancillary stuff further down the page, whether that's get preapproved, do a test drive, ask a question, and we really try to lead with urgency and things that actually appeal to the customers.

We're constantly adjusting that. We're looking at the heatmap data and analytics and we're looking at session recordings to see how people interact with the sites.

It's a little different for every dealership. But I think the underlying principles are the same. Show them what they want to see on something relevant and get them into the funnel. And interestingly, for some of the higher-end dealerships, Fisher Imports here in town, they were one of the first clients whom we deployed our retailing tools with. When we put "explore buying options", they got quite a bit of clicks like you could go through the whole process of your trading and your finance forum, all of it. But once we actually just added put down a deposit, that's one of his top performing conversion mechanisms and that requires somebody to put down a deposit with a credit card. It's not a free hold or anything like that, but for people that are shopping and know what they want, they're willing to put down $250 or $500 to, you know, lock into the car they want.

Can you share some interesting website analytics?

I've got a few. A lot of the leads that we see tend to visit two or three cars. It just really depends on the source of the traffic. A lot of dealerships are investing in, you know, blogging or content generation. That's honestly, the poorest performing. We see paid ads are sort of in the middle somewhere. It just depends on what are they collecting, what they are doing.

But overwhelmingly, I would say the best-converting traffic we see is organic from search engines. We're starting to see that the best-performing sites are typically 60 to 70% organic traffic. It's really sort of changing dealerships' behavior because rather than spending a ton on pay-per-click or referral websites, we see them investing more on the SEO and search engine side because it performs a lot better.

I have a funny statistic for you since we're on the topic of chat earlier. When we look at heat map reports on a vehicle detail page, the number one click is the close button on live chat and it accounts for nearly 21% of all clicks. 

There are some other things we track. Like what people are searching by. We track features. Features are really interesting because they seasonally change. So in the summertime, people are primarily clicking Apple CarPlay sunroof, all these things that are conducive to the nice weather. And then as soon as winter hits, it's all like all wheel drive, towing package, or whatever. It's all things that are sort of gearing up for the winter. So that's always interesting to see.

What trends are you seeing?

My business partner pays attention to search trends quite a bit. We're seeing about a 20% decline in used car searches which is interesting. If you go to Google search trends and put in used cars like that, numbers are trending down. So I think that that's pretty interesting.

Another one would be pay-per-click. So when I was talking with Alex Melon about predictions for this year, we were looking at some of our numbers and wanted to compare them to some of his numbers. And we see a handful of dealerships that were paying a dollar per click back in January of last year, they were all of a sudden paying three plus dollars for that saving click at the end of December. So paper click costs are just multiplying every year and it's just getting so much more competitive.

Site speed and search 

I think automotive has struggled. You talked about speed a little bit, but the speed benchmark that we looked at said only 39% of automotive websites actually meet Google's sort of core performance metrics. Only 39% of automotive websites are tuned to perform well on organic search or paid search. It is the second worst-rated industry, only second to travel. I think historically, the industry has really struggled from a speed standpoint. And for clients that we see invest in proper SEO and a great mobile experience, mobile-first speed, we see increases of 2 to 300% in organic traffic. And to your point, organic traffic performs better so that often ends up being a 2 to 3 X increase in leads. I've yet to see a single pay-per-click agency or an SEO agency deliver that kind of performance because ultimately we have to solve what's under the hood, like what's causing the performance issues in the first place.

What is slowing down dealer websites?

It's a number of things. There's a metric called total blocking time, and that is just the time that it takes for an end user to connect to your website. That is directly in the hands of the website providers, whether they're using highly available infrastructure and cloud hosting and all these things they can do to deliver the website to customers based on their location to try to get a closer proximity to the data center. That's just right off the top. The most obvious one. If you're like a lot of website vendors in the space that white-label a content management system like WordPress, you're not going to be able to optimize that metric very well because WordPress is not built for automotive websites. It's got all these other plug-ins and all these other configurations that it has to load before a customer can even see your site. So that's number one. That's a design decision on the software side of things. 

There are two other major contributors that we see very often. So a lot of dealership websites today are not efficiently serving up images or videos or assets. We've seen the dealerships that have 25-megabyte homepages. So like to compare that to one of ours. We're like five or below. They have these huge promotional banners and really large videos, and you have to think about how that's going to load on some 3G cell phone, right? It's not going to go well. 

And then the third really big one is just the fact that people aren't closely paying attention to the widgets that are being put on their website. When we migrate a dealership over, a couple of the activities that we do is we try to consolidate all of their Google Analytics accounts and we try to consolidate older Google tag number tags. I think the record holder that we moved over had something like six analytics tags and 12 Google Tag Managers. And that alone starts adding seconds to your load time.

It's really just because you had a vendor that had you add this tag or you had an advertising agency that wanted to add a tag. And most support teams or customer success teams aren't considering the implications of that. So one of the things we monitor very closely is how many new scripts and new tags have been added to the website so that we can go back and empirically point to that and say, Well, this is when things started to go south and then we figure out how to consolidate them and and keep it performing well.

Google Tag Manager is slowing sites down.

Well, Google's being kind of quiet as to why that happened. I mean, that hasn't always been the case. I would say it's a fairly recent development, maybe the last 3 to 6 months that the tag managers especially are performing really poorly. If you go to the Google developer forums, there are a lot of people asking what is going on. And the conversations are just closed, unresolved and they're not allowing new people to comment. So it's kind of unknown at this point as to when they're going to fix it. So the best thing we could do as website providers is to try to optimize around that,

How are website providers contributing to slow sites?

It's I mean, it's a good question. I think if you went and looked at the list of industry leaders, the first thing that stands out is there are some very dated technology choices being made. I empathize with that because I'm the last person who wants to go back and rebuild my entire infrastructure. But I just think everything's evolved so rapidly. You have modern JavaScript frameworks now that are powering services like Hulu and Twitch and Netflix, and it's really hard for a company that's ten or 15 or 20 years old to adapt that kind of technology and get it implemented.

It can be very challenging. That is the software or the engineering culture in the industry. The dealers keep paying and we don't have to innovate and people accept the status quo. We talked to a very sizable company in the space that uses the KATL acronym. Keep on the lights. That seems to be one of the big challenges that we've noticed. They're just keeping the lights on instead of trying to innovate because they haven't been pushed or demanded to innovate in a long time.

What can a dealer do to improve their website today?

A couple of things that we often do is we look at a tool called PageSpeed and any dealer can go in and put their website in and PageSpeed will tell them exactly what's going south.

Another good one is GTmetrix, and any dealer can go in and put their site into one of these tools. It could be a little complicated. You know, it kind of gets a little nerdy, but you know GTmetrix especially is very good about sort of showing like this little ladder of how everything is loading and what's taking forever. They can pretty quickly see, "why did my marketing team add ten tag managers to the site?" It really spells out what the issues are and in a friendly matter, it even gives it either a color-coded score or a color-coded grade. So that's an easy one.

One that's really low-hanging fruit, if you want to understand if you have a good SEO vendor or a good SEO strategy in place, I highly recommend that dealers go to a tool called Semrush and they can put their business name in their directory listings management tool and it will show okay, of the 50 plus trusted by Google, here's how many your set up for. That's a really simple feedback mechanism. You have 50 websites that Google trusts that will index or website, make sure that all your name, address, and phone information is correct and it broadcasts that out to the Internet. That is the most low-hanging fruit thing any dealership can do. And nine times out of ten, when we talk to a dealership that's paying an eye SEO firm., that's the first thing we check, and none of their directory listings are on par. That's costing them a lot of organic traffic and really calls them a lot of opportunity.

What can a dealer do if they have an underperforming website?

I would agree that they're probably just going to have to find one to do the job. Unfortunately, there aren't very many right now, but you know, you brought up DMS, it's like we saw Tekion come into the space and they're working on some pretty exciting stuff.

There are more people sort of coming into the space wanting to do things a little differently. It's just unfortunate that we're also met with a lot of reluctance because as I look at it, you're the 50th DMS provider or the 50th website provider to call me. Why are you different? And so that's going to be an uphill battle for any company that comes into the space and tries to make an impact.

But what I will say is that if you have a great product and you can prove that it works the dealerships are your best friend in terms of marketing. If one guy starts doing really well they tell the next guy. And so that's just where I think the industry struggles, because I'm not sure what word you would use to describe it, but it's almost like dealerships feel a little defeated because they keep investing all these dollars expecting a different outcome. And I feel like they've been kind of let down for a while.

How do you approach digital retailing?

Digital retailing is a very interesting topic because it's kind of a kind of a loaded term. I think at the very basics, you should have some true payment calculation, some sort of input so the customer can put in to get their real payment. That way it's transparent because what we've seen with a lot of retail tools is that they present pricing with the best possible terms and credit rating and all of the above. Then the customer gets sticker shock when they find out the real price.

I think where a lot of retailing tools sort of overcomplicate things is that they try to package the entire deal into a single workflow, and for us that has been the lowest-performing plug-in across the board, not only for ours, we have our own retailing plugin, but we typically see it other retailing plugins fall short because they're just too complicated for the attention span of the modern consumer. That's where I was going back to, if you have something more actionable that's like, Hey, put down a deposit on a car and reserve it for 72 hours. And if it doesn't work out, it doesn't work out. It's much easier to convince someone to do that than it is to get them to go through a five-step credit application and then a three-step trade-in form.

Ultimately it's still a gray area as to whether somebody could do all of those pieces of the purchasing puzzle and actually complete a transaction. I mean, more often than not, what we see is that somebody will put down a deposit, the dealership reaches out, what are my next steps? So they get instructed to put in a credit application. They talk through the numbers over the phone and some of our dealerships ship 50% of their cars out of state. I think the expectation that you're going to wrap all of that into a single experience is going to be very challenging. 

Caravan and Vroom do that particularly well. I think it's a general generational thing. I think they appeal to younger car buyers. But I would think the average consumer, when you're making a purchase decision like that, you still want to talk to somebody on the phone. And part of the retail pitch that I've heard is, to your point earlier, is bring up more man hours or free up more resources. But you know, somebody is going to spend 50, 60, $100,000 on the car. Just give them the option to reserve it and sort of take it from there.

Digital Retailing needs to be simple

When dealerships often ask, do you have digital retailing tools, our answer is often yes. Carvana and Vroom and all these guys already have proven that people are willing to put the money down and buy a car online, whether that's the entire process or they're just putting down a deposit and getting started. But we know the desire is there for consumers to be able to transact online.

I think where dealerships get in trouble is relying on all of these third-party plug-ins with dozens of startups with this false promise that it's going to generate a ton of leads. We made that same mistake. We went down that road. Ultimately just making it one simple step seems to be the most effective.

How are dealerships calculating ROI on search activities?

You made a point earlier about dealerships looking at their budgets and trying to figure out how they quantify results and how they should allocate their dollars. I would tell you that the vast majority, 95% plus of dealerships that we talk to can't actually calculate the the ROI or the return on investment from different search activities or advertising activities. 

In large part that's because every time a customer comes to their website, they are sort of shuffled between a live chat product or an embedded credit app or some sort of instant cash trade offer. Once those people engage with those tools, you no longer have any visibility into what they did. So the reporting for dealerships is extremely challenging.

That's one thing that we're sort of aiming to solve. We have our own proprietary analytics engine that tracks every single interaction, every single detail. that's kind of the visibility that you need now because even Google Analytics has taken a huge nosedive with the release of Google Analytics. All the data is delayed 24 hours. The retention of that data is very limited. So you can't really compare time periods and what performed or didn't perform well.

They're really interested in selling looker subscriptions with analytics. It's not quite the tool that it used to be for generating reports. So it's very challenging across the board. As I said, probably 95% plus of the dealerships we talk to often couldn't tell you what their most effective marketing spend is.