How Hewlett Packard Enterprise Personalizes Their Website Experience
Bill Mitchell, Manager of Digital Demand Generation at Hewlett Packard Enterprise, has over two decades of experience in digital marketing. On the B2B Growth Podcast with James Carbary, Bill talks about his team’s experience growing website personalization from a simple experiment on the homepage to a suite of programs that directly influence revenue.
James Carbary: Welcome back to the B2B Growth show. We are here today with Bill Mitchell. He is the manager of Digital Demand Generation at Hewlett Packard Enterprise. Bill, how you doing today?
Bill Mitchell: I’m doing great, James. Thanks for having me.
James Carbary: I am really excited to chat with you today, Bill. We’re going to be talking about, really, personalization and Hewlett Packard’s journey that you guys have been on for, I think you said, the last three years is when you guys really started focusing in on personalization?
Bill Mitchell: That’s right. It might even go back further than that, but we really started to make our inroads about three years ago.
James Carbary: I love it. Before we get into sharing this journey and what you guys have learned along the way, I’d love for you to just give a little bit more context as to—I think everyone listening to this is familiar with Hewlett Packard—but talk to us a little bit more about your role at the company, and what the company is trying to accomplish.
Bill Mitchell: Hewlett Packard Enterprises is the result of the spin off from HP a couple of years back now. It’s a two year old company in some sense, but we also have nearly 80 years of history and heritage to lean on. The split was something that the company went through, but it also from a digital marketing perspective, provided the HPE side of the house an opportunity to do a lot of things differently. We’ve made a lot of enhancements to our digital practice. We had to reset our foundation in terms of the tools we use. We also had to put forth a little more focus on how we wanted to digitally execute, so it’s been an exciting time these last two years as the company has started to find its groove as we moved into a more digital-centric marketing play. One role is obviously to support the enablement of that and working with the website. How we develop out functionality that supports all sorts of programs, personalization being a big one of those, and then some of the capabilities of actually driving those programs. I lead two groups that are building things that support the site, and also driving some of the programs that are developing the content and execution on the site.
James Carbary: Bill, talk to us about when you guys first started looking at personalization, before starting to work with Bound. What were the things that lead you to making that decision — “Hey, we need to start getting more proactive in the way that we personalize the site.”
Bill Mitchell: That’s the optimal question, is how does that journey get started? Going back three years ago, we were, and we still are, a very complex company. It’s very hard for someone our size with lots of products, lots of different visitors, lots of different needs for making your way to our site. What are we trying to share with somebody? When you do that in a default manner, it’s very hard to ever be right with the guesses you’re making with the content that you’re hand curating to make available. The initial inertia around this was how do we put forth—it started with the home page, much like probably everybody else who has done personalization—how do we at least start to tailor our homepage to cater to folks that we can recognize, to either make a product recommendation, or guide them towards the corridor, or things that would be more appropriate for them, rather than just the default experience, which hits the mark probably almost never.
James Carbary: You make this decision that you can’t keep depending on just the guesses that you guys were making, and so talk to us about the decision process and how you landed on choosing Bound to be the partner, and then from there I want to get into the nitty gritty of the story of what you started to personalize and the results that you started to see. As far as choosing a vendor and choosing someone that you knew you wanted to partner with on this initiative. Walk us through what that looked like for you.
Bill Mitchell: At the time I was in our software business unit, which in and of itself was a pretty large substantial footprint. What we wanted to do is find some tools that would allow us to manage that program right from our business unit. We were attached to a corporate entity at the time that had its own tool set for how they delivered different capabilities, but they weren’t very conducive for us to start making agile smart decisions that were tailored to what we needed to accomplish. So when we were looking for a tool, we needed something that was easy to get off the ground, insert a pixel and off and running type of capability. Something that was easy to maintain with very little handholding. We didn’t have an IT group that would support setting up tests or configuring things on our behalf, so we were looking for something that was easy to deploy, easy to use, and would allow us to get a pretty quick return on our investment as we’re able to stand up a simple model and see how it was returning value to us.
James Carbary: So you guys ultimately decided to use Bound, and in doing so, as we were talking offline, you mentioned that the first type of personalization that you started to focus on was on your homepage. Talk to us a little bit about what you guys were personalizing there and what the results were from that personalization.
Bill Mitchell: I can’t imagine we’re that much different than anyone else, like I said, that’s been on this journey, but we picked a premiere spot on our homepage, and what we attempted to do at the time was to inject replacement content based on some simple rules that we had set up around different industries, and if we had seen return visitors, obviously we could then tailor it a little better based on what they had done in a prior visit. So we carved out that real estate on the page. We built out some different content options that would be replaced based on whether or not someone triggered those different rules. So a return visitor or someone that came in from a particular industry, would see now something different. So we were starting to tailor the page, a minor part of the page, visible but minor in the sense that there’s maybe a hundred things to click on on a page and we were tailoring one of them to those individuals that were visiting our site that we could drop into a segment.
At the time that was extremely beneficial. The results we were seeing there were triple digits. The number of people clicking on default content versus the personalized content was well over 100%. Depending on what segment and what piece of content we were seeing, sometimes near 200%. The early findings were very helpful to help us keep that program moving along.
James Carbary: You mentioned this earlier, Bill. You said that the criteria for what triggered the personalization — one of them was have they visited the site before. What were some of the other criteria that were triggering the types of content that they would see?
Bill Mitchell: Beyond that initially we were looking at previous visit and industry, and industry is based on the IP of the visitor to the page, so if we could detect the IP and use that insight to drive a company or an industry, we were using those to help us create those segments that our content would be married up against.
James Carbary: You obviously saw significant results there. But that’s not the end of the journey. You guys saw results there, and then you wanted to take it another step further. Talk to us about the next evolution in the journey beyond the results you saw from the homepage.
Bill Mitchell: The homepage, like I said, we were seeing great results. They’re limited, ’cause there’s lots of things to do on a homepage, and even though that particular content element was thriving, we could go on a path where we were updating dozens of content elements on our homepage, and trying to increase the impressions and the footprint of personalized approach. What we decided to do is let’s go into one of our most popular product pages on the site. An area where we already know what people are interested in. They’ve already predetermined that based on their showing up on that particular page. But how can we make the content of that page more appropriate for them? This was a pretty cool, at the time anyway, concept—where if it was a new person or someone that had not done anything to show engagement yet, we isolated early stage content for them.
We mapped our content along the buyer’s journey, and what we attempted to do was to isolate the visitors to the site into three buckets. Early stage awareness, browsing type of individual. Someone that’s returning to the page or in that visit is showing active engagement by clicking on various things. Or the later stage folks, that is have clicked on and even maybe converted, so at that stage those types of individuals were looking to pull them closer to sales and showing different types of content to get them to reveal themselves, or to raise a hand and show that they have sales intent in their particular visit.
Those three categories is what we broke our content around. Click that one page and we’re drastically changing it to provide those types of content options for folks based on those predetermined segments.
James Carbary: Talk to us about the results that you saw from that next evolution of personalization.
Bill Mitchell: Now we’re personalizing multiple spots on the page, and it started to get—and this is probably maybe six months into our efforts—and Bound had released a new capability and let us fly content into the page, which was a really attractive way for us to start to present content that didn’t have to be injected into the page. You’re dealing with Flicker and you have to have a massive backlog of content that would support such a type of path, but with this new capability—and it’s not new anymore—but at the time this new capability would let us create a series of promos, if you will, almost banners that were literally floating into the page. And those types of content were getting, on this particular page, some of them were getting 2% to 4% clicks. Some as high as small double digits.
Cumulatively now I’ve got content flying into the page with those different segments, and we’re starting to gather clicks both on that page, and we’re also able to take those same fly in offers and present them on any page in our site. So when I knew someone had interest in that product, in that visit or a returning visit, if they were to stay on our homepage, I could fly in an offer back to that particular piece of content, so allow us to extend the footprint of that page across our entire site once someone had stepped into one of those segments. Those 2% to 4%, or 8% to 14% types of clicks we were getting, were occurring on the homepage, but basically crediting a click back to that particular product.
Now we’re getting to a scale model which was really encouraging, and it led us to really appreciate how well chat was doing for us. Chat was something that we made available, but had included now as one of those content elements, and we’re seeing outstanding results for it, which kind of triggered the next generation of our approach where we took chat and started to run with it.
James Carbary: Just to camp out here, ’cause I want to make sure I’m understanding this, and our listeners are understanding it. The fly in functionality that Bound released. Whenever a user would come to your site they would click through to one of the product pages, and then maybe they would leave. When they come back, Bound essentially allowed you guys to bring a piece of content to the homepage of that next visit that would essentially draw them back to the deeper element of the product that they had looked at before. Am I understanding that right?
Bill Mitchell: Absolutely, yeah. That’s exactly the model. What we were doing is trying to lead people through the buyer’s journey, and we could have continued to do that on that particular page, and we did quite frankly, but it also now with this fly in approach, I didn’t have to manipulate content on each and every page. I could fly in over the page, and that gave us a lot better opportunity to get some scale, because now any page in our site where we recognize those visitors was an opportunity to extend content to them, and again, they’ve shown interest in that particular product. They’ve identified themselves in the model that we set up of where they sit in that journey, so we were able to go into that bucket of content and make an offer to them to try to pull them back to that initial product where they’d shown that early interest.
James Carbary: So each time that they come back to the site, it puts another piece of content in front of them that pushes them a little bit further down that journey?
Bill Mitchell: Yes. That’s the approach. That now we’re starting to see great results from. I’m already taking someone that’s shown interest in that product. I’m now breaking my content into these three buckets where I’m either with you if you’re still in awareness, I’m going to present content to you that’s helping you get more aware about what that product has to offer; I’m either trying to get you to engage now and move you further into the buyer’s journey; or I’m trying to get you into a sales motion as you’ve shown further interest, and we start to extend different offers that are trying to pull you from left to right, if you will, if you’re looking at a linear journey.
James Carbary: That’s super interesting to me, but now I want to—as we end the interview today, I want to talk about this—the way that you guys have been able to personalize the invitation to get users on your site to engage with the Chat tool on your site. Talk to us about that. You alluded to it before, but I really want to go deep here, because I think this can be really powerful.
Bill Mitchell: One of the findings we saw out of that product page approach is the offers we made. One of them was Chat. It was by far surpassing what we were getting from Chat out of the box with the tools we were using at the time. Double, triple the amount of clicks we were seeing, which quickly led us to say “Why aren’t we doing this in more places? How do we capitalize on this more effectively?” We quickly put together a program around how do we extend chat. Essentially turned off the capability in the default tool that was extending the offers, and we moved to Bound to deliver that capability.
It blew the doors off what we were seeing in the past. We were seeing 200% to 300% more Chats based on the invitations we were extending. So we were still using that fly in approach, no different than the default tool. But now with our abilities to extend the reach across the whole site, as well as personalize it based on some various simple factors, we had the agent say “Can I speak to you about this industry?” “Welcome back to the site. Would you like to chat with us again?” All these types of things that were communicating in a much more specific way. We were, like I said, seeing 200% to 300% more clicks, and we doubled the number of chats that actually ended up as leads with our pre-sale seeds. That at the end of the day started to become, that’s the moment for us where personalization was no longer nice to have or a cute side project. It was now leading to direct revenue that I could attribute back to our abilities to personalize and to run that program at scale.
James Carbary: It’s incredible. You were personalizing that invitation to chat in similar ways that you were doing with the homepage — you’re doing it by industry, or by something that says “Welcome back.”, indicating that we know this isn’t the first time you’ve been here. Those are two of the big ones. Were there any other ways that you were able to personalize that chat box that you saw significant uptake in, or were those the two primary ones?
Bill Mitchell: There’s a third one we had now that … where probably 18 months ago where Bound had built a relationship with LinkedIn, which led us appreciate where visitors to our site that had LinkedIn accounts, that we could trace back to their roles, which now we’ve opened up the door to someone’s function at work. What kind of role they have. It’s not titles or anything. PII-oriented per se, but it gave us some insights into what area of the company that person worked in. So then we looked at what are the biggest footprint of people on our site from those LinkedIn profiles, and what types of job roles did they have? So we even built out chat invitations that would speak to you as someone in the QA department or someone that has worked in the marketing function or something like that. Messages again, that we’re trying to tailor, in this case, still chat, but using the messaging in a much more one-to-one manner.
James Carbary: This has got my brain racing, Bill, on just different ways that you can capitalize on personalization. I love this last piece of the tangible results that you guys have seen as a result of customizing each chat window. If I’m picturing this right, it pops us in the bottom right corner or whatever it would be. The ones that I’ve seen — it’s almost like an agent is starting the conversation with you, so it’s just that first message that the person on the other end of the chat box is sending, and you guys played with what, probably 15 to 20 different messages that popped up in that initial message?
Bill Mitchell: Yeah, that was actually— thanks for reminding me. That’s another thing that Bound allowed us to do. Because now we’re using not just a personalization tool, but it has testing capabilities as well, so we weren’t just sending out one message per segment and resting on our laurels. We’re actually able to test different messages, different copy, different creative. Should we use our company logo or a picture of an agent or some other type of creative to actually help these invitations stand above the fold? So yeah, you’re right. It was just a simple fall into the page I think on the lower left or what have you, but manipulating the creative in multiple different iterations gave us even more learnings and helped us optimize the program.
So, for a particular segment we have probably two recipes running at any point in time, and as we saw which one was drawing better results, we were able to solidify around that, or even in many cases, roll that particular winner into a new test where we try to extend even further to see again how far can we push. What is the maximum number of chats we can get from these given audiences when we have the right timing, the right creative, the right messages, all those things working in harmony?
James Carbary: What did you guys find, Bill, I’m just curious, from a creative standpoint? Did you find that an agent’s face worked better? Was there a particular message that ended up working better for you that you’d be able to share?
Bill Mitchell: Yeah, it’s funny you mention that. The three things that we were really rallying around was a picture of an agent, which looked like a picture of a person. A picture of an agent which looked more like an illustration, though a person but just an outline, a drawing if you will. And our company logo. Our initial bets were the picture of a human would draw better, being faces tend to draw better in terms of clicks. It turned out our logo worked better than either of those two agent-based perspectives. Obviously we work at a pretty powerful brand, so there’s something about that, that probably skewed things. But it was very interesting because our initial impressions in what we thought would be most successful turned out not to be.
James Carbary: Interesting. The other thing that you mentioned that I love that you talked about, is you guys didn’t just land on a winner and camp out there. You guys are continuing to A/B test, taking those winners, rolling those winners into new tests. Is that something you guys are doing on a weekly basis? Are you running these different A/B tests quarterly? What’s that look like as far as the cadence of how often you guys are testing?
Bill Mitchell: Our program is very intertwined with testing, and has been, not since the beginning, but pretty early on. We do testing in isolation of personalization, but you’re typically using the same tool, so it’s a smart place to not just personalize the page, but to test what you’re personalizing, and vice versa. You can use personalization to manage some of the testing. We run them kind of always on, and the timing will differ depending on how quickly you can get to statistical significance. Things like Chat where it’s hitting virtually every visitor to the site, we have pretty quick learnings. You’ll know in days to a week where your winners are. Again, depending on how many segments you have and how many people show up in those various segments. We may have a default segment where you learn in two days.
You may have a micro segment around a particular industry tied to a return visitor where it could take you two months to learn what’s the win recipe just based on volumes. It’s ongoing. We’re constantly adding and removing things as we see winners appear, and as we have new ideas to test the creative. Constantly challenging that we can’t assume we know what’s best, we have to keep testing to prove what’s best.
James Carbary: I love that mindset of continually testing. Is it something that you’ve empowered your entire team built to say “Hey I’ll always be thinking about different ways that we can test and iterate and move.” Or is it do you guys have a weekly stand up, or where do you see a lot of those ideas for different types of creative popping up?
Bill Mitchell: We have both. We definitely have a weekly or biweekly cadence where our teams will get together to look at results, and to explore new opportunities. But it’s also ad hoc as well. Literally today, I had a conversation through email with a few folks around a different idea, and it happened to be with chat actually. And I floated the idea by our chat admin, who’s leading our personalization around that programming. Today, this morning, it was kind of funny, the timing here. But he was working up a couple of examples that would allow us to explore that different idea. So we definitely have the methodical cadence of how we manage the program and look at testing opportunities, but you never know where or who ideas will come from, and we’re always looking to jump on those and take advantage of the learnings. I want my team to always be looking at ways to improve our performance, and we’re pretty data driven around here, so that I think approach leads to people looking for ideas, looking for new opportunities, and always striving for the next best recipe that’s going to push us further towards our goals.
James Carbary: Bill, this has been fantastic. I’ve learned a ton. I know our listeners are going to get a ton out of this interview as well. If there’s somebody out there listening that want to stay connected with you, they want to learn more about HPE, what’s the best way for them to go about doing both of those things?
Bill Mitchell: To learn about the company, definitely check us out at HPE.com. I encourage you to check us out from like a mobile device and maybe on your work computer, and see if you can appreciate some of the differences that our occurring, depending on your industry and various things. I wouldn’t be surprised if you notice looking at our site between your desktop and your phone, you’ll most likely see some differences as you scroll through our site. To reach out to me I can be found on LinkedIn. My LinkedIn name is GW Mitchell. There’s lots of Bill Mitchells, but I think I’m the only GW Mitchell, or G William, sorry. You can find me there.
James Carbary: Wonderful, and if you’re interested in Bound at all you can go to Bound360.com and check out all the solutions that we’ve been talking about today with Bill. They have been using Bound for the last three years, to what seems like really great success. So go to Bound360.com to check that out. Bill, thank you so much for your time today. This has been fantastic. I really appreciate it.
Bill Mitchell: Thank you. Thanks again for having me.