Spotted: Web personalization for B2B

Websites are being left behind in a data-driven, real-time personalization world. When email, ads, live chat, sales enablement, and other channels are able to leverage lead & customer data to deeply personalize the way they engage, most websites show the same experience to everyone.

Why is the web left last — despite having some of the biggest, easiest marketing reach?

And how are data-driven teams managing to personalize their web experience?

In this Spotted post, we explore the tactics and strategies we’ve noticed data-driven teams using to drive real-time web personalization for B2B.

We went from one average best homepage to 26 different versions of the homepage uniquely personalized to different visitors.

Cara HarshamnOptimizely

Personalization works ontop of A/B testing technology.

A/B testing technology became popular in SaaS since 2010 with the birth of VWO and Optimizely (who offered a free plan in their early years). These technologies made it relatively seamless to edit and track different content experiences for different audiences - the primary use case to manage A/B testing.

But it wasn’t just useful A/B testing. You could create audiences to display different messages (like CTAs) to different audiences. Responsive design had been widely available since 2008 — now we could create responsive content for different website visitors.

Audiences in A/B testing tools could be created from browser behavior, but this is just a fraction of all the data known about a particular person. To unlock deeper levels of personalization, you need to tie your tools to profile data (like you do for other channels). For B2B

Meaningful personalizing for the web is hard without profile data

Email, CRM, ads, live chat, and other channels (where you can personalize content easier) work around a profile — this sits on the server side.

But web personalization works client side “in the page” — its only record of a person is the website visitor it has in front of it. Almost exclusively, this depends on whatever data is in the browser client. IP, location, previous cookies (indicating repeat visitors), and so on. There’s still plenty of segmentation that’s possible, but not to the same degree you can personalize email, live chat, and your sales calls.

In B2B, there’s a much more significant gap. Since you’re selling to a company and everyone in the company, the browser behavior of just one person from a company doesn’t give the full picture and context you need to deeply personalize your content.

  • Are they a good-fit customer? Right size? Right industry?
  • Or are they already a customer? Partner? Competitor? Are they already in the sales cycle?
  • Have other people visited your website from the same company?
  • Are people from the company engaging with your brand in other channels? Email? Ads? Chat? Sales?

web-personalization-1st-degree-loop-optimizely

The limitation of data is the limitation of the content — without profile data in the page, there’s only so much you can do to precisely personalize your web experience. The challenge is how to get data into the page - quickly!

Page flicker: the bane of web personalization

Page flicker

Even with data able to get into the page, it has milliseconds to reach your personalization software. The team at WiderFunnel have a great write up on why page flicker happens and what to do about it.

Customer data integration between other tools has the luxury of time (seconds-to-minutes), whereas a website has to render a page in milliseconds to avoid a website visitor bouncing or experiencing the “flicker” as the original set of content loaded is swapped for personalized content. Few teams and tools can push all the profile data they might need to personalize into the page fast enough.

But this hasn’t stopped some teams. The lure of the lift from getting personalization right in their biggest channel has driven teams to explore the limits of data integration, data processing, content serving, and artificial intelligence to replicate and iterate on their personalization programs for the web.

Now I want to go further. I want to be able to collect, analyze, and act across all channels — not just chat, not just email, not just the webpage.

Guillaume CabaneVP Growth at Drift

Here’s the five tactics we see powering this web personalization trend amongst B2B teams.

Tactical Takeaway #1: Move beyond browser behavior for B2B

There is more to a person’s customer journey than what is available by default in the browser.

For B2B, we seems working to move beyond just browser behavior. Compared to B2C brands, there’s far more accessible, depth of data to deploy from your CRM, marketing automation & email, ads, 3rd party data enrich sources, live chat conversations, and even your backend database.

Segmenting just by location or returning visitor may be helpful, but you it isn't as meaningful personalization as what you might be seeing and using in your other cahnnels

Voila! You’re no longer constrained by web as a channel compared to your other channels.

Tactical Takeaway #2: Identify companies who are visiting your website

The first step is associate the website visitor with a company — without requiring their email address.

Optimizely themselves shared how they personalized their own homepage to drive a 117% increase in accounts created.

You can identify businesses based on the IP address of the website visitor since most businesses use office space with a common IP address.

Reverse IP lookup services can transform an IP address to a company name. Though Optimizely used Demandbase, we noticed most B2B SaaS companies avoid Demandbase — they have pricing where you pay for each tool where you used your data. (At Hull, we’re not a fan. We believe in breaking, not making data silos).

Clearbit Reveal is a popular reverse IP lookup solution for B2B SaaS companies that lets you use your data everywhere (like all your tools through Hull). The team’s we’ve talked to found better match rates (particularly amongst mid-market target accounts, not just identifying big enterprise companies).

You can test Reveal data out in your Google Analytics — it looks something like this.

Clearbit Reveal Google Analytics

Voila! You now know which companies are on your website

Tactical Takeaway #3: Use enriched data via a customer data platform to power web content variations

Google Analytics is useful for retrospective analysis, but not real-time action. You need to be able to sync the response from your reverse IP lookup service back to your A/B testing & experimentation tools.

In the case of Clearbit Reveal, the response also returns the full Clearbit Company payload with 40+ data points on the companies size, industry, location, and more. These can all be used to segment website visitors and display different content variations.

However, reverse IP lookup can also associate data with everything else at account-level. This is where a unified customer profile and unified company account profile within a customer data platform holds a unique advantage for web personalization. Instead of querying dozens of different places to respond in real-time (to load a content variation), you can just query one place.

Given the high volume of requests web traffic generates, a customer data platform is much better placed to handle this compared to your data warehouse, CRM (Salesforce is limited to 15,000 API calls per day, then you start paying substantially more), or marketing automation tool which holds the rest of the customer profile data.

Using the Hull Browser connector with Fast Lane enabled (for fast-tracking processing of “live” web traffic), you can sync all or anything from your unified account profiles into the browser client to power personalization in one call. Collectively, this gives unmatched depth of data to personalize your web experience.

Measure fit with;

  • Clearbit Company data
  • CRM notes
  • Lifecycle stage

Measure account-level intent with:

  • Browser behavior
  • Previous live chat conversations
  • Previous email engagement
  • Previous website visits (with an enduring stable domain)
  • Previous product usage
  • Previous social media clicks
  • Previous ads clicked

… from any person associated with that account.

Web personalization 3rd degree loop

Voila! Now you have real-time data to power meaningful web personalization for B2B

Tactical Takeaway #4: Manage audiences & variations with predictive personalization tools

It’s is much easier to personalize well defined, constrained content formats like email, live chat, and ads. But teams who have designed & implemented any substantial web development or A/B testing projects know it is much harder to manage websites — with fewer boundaries comes more scope for complexity.

Tools like RightMessage can make this quite seamless to setup (or you can implement in your existing A/B testing tool). There’s are some early wins like homepage variations, signup flows, and pricing pages.

But to move beyond this and test across more pages and parts of the your content, you might running

  • Managing audiences
  • Managing personalized variations
  • Choosing (or guessing) winning combinations of audiences and variations for every page.

Away from the early wins which likely have all the traffic, it becomes hard and time consuming to setup winning personalized variations. Since most B2B sites lack the traffic of B2C, management of personalization becomes a barrier to experimentation.

This is why we see data-driven teams implement B2B personalization at scale through predictive personalization platforms which automatically match audiences and variations instead of marketers managing them manually. This enables them to dig beyond the low hanging fruit and deliver sitewide personalization.

A/B Testing gave us a way to understand the best single experience for all visitors at one moment in time, and predictive personalization now gives us a way to deliver the best experiences for each and every visitor as visitor behavior changes.

Guy YalifCEO at Intellimize

The choice tool we’ve seen used is Intellimize. Instead of creating variations and assigning traffic, Intellimize chooses which variations are shown to which visitor, measures the engagement, then updates a predictive model to reallocate types of traffic to specific combinations of winning content. What?

In short, Intellimize manages multivariate testing down to a 1:1 audience for you.

This means you can confidently test and experiment with the content assigned to each segment since the legwork of matchmaking is handled for you. This means you can experiment really fast. Intellimize was used on Segment.com to test 105 ideas across 4.5 billion (with a b!) personalization variations of their homepage in 3 months.

Like traditional A/B testing tools, Intellimize works in the browser. The same challenge of real-time data integration exists. For B2B, reverse IP lookup and using that to query Hull to stream the response back to Intellimize via Hull Browser. Intellimize will update its predictions based on whatever data is available. This is the system that makes B2B web personalization possible.

Voila! You can now have a system to manage all the complexity of B2B web personalization as you scale up your web program.

More "Spotted" playbooks like this?

Subscribe to The Crow's Nest newsletter from Hull for best practices, ideas & resources for customer data management.

Tactical Takeaway #5: Matchup your web personalization with experiences in other channels

At Hull, we believe in breaking silos, not making silos. Your hyper-personalized web content experience should follow through with the rest of your channels too.

The volume you can drive through the web is likely to be significantly greater than your other channels too — this trains predictive models much faster and more accurately than otherwise. So try enriching other tools with the insights you’ve discovered from your web personalization experiments.

To do this, you need to track which variation has been viewed to which person. This can be tracked as an event with event properties. Here’s an example of what that looks like in Hull…

// Track Intellimize variation as event property

hull.track("Viewed Intellimize Variation", {
  Experience: <value>,
  Audience: <value>
})

Use event-based segmentation and event property-based segmentation to segment your audiences based on the experiences and variations each website visitor has viewed. You can create event property based segments with Hull’s segmentation tool.

Use these segments follow through and follow up with your messaging, or to introduce “new” missing or supplementary messaging that wasn’t covered in the original web content.

We're looking at acquisition data (original landing page, referrer) and then on-site behavior (type of content), grouping it into segments (e.g. "domains read by designers") to increase top-of-funnel opt-ins & syncing that segment to our CRM and email content

Brennan DunnCo-Founder at RightMessage

The first place this makes sense to use is within website live chat. Since this also works in the page, personalizing live chat ontop of your real-time web brings a harmonious web experience.

Since live chat doesn’t have to load right away, live chat also has a little more time to fetch more & richer data. You need to choose a live chat tool that enables real-time data to trigger & personalize messages to website visitors. At the time of writing, you cannot trigger new messages through Intercom’s API to visitors (yet!), but you can with Drift’s.

You can also use the experience from personalized web to update your email lists, ad audiences, and write in activities into your CRM like any other tool. Checkout how we’ve spotted other teams managing personalized, omnichannel lead nurturing.

Voila! You’re now able to orchestrate personalized content across web, email, chat, sales — everywhere!

Best-fit criteria for real-time web personalization for B2B

Real-time personalization isn’t for everyone. We recommend you lay this on top of a website that already has plenty of traffic and converts. That in itself can be a significant challenge in B2B — particularly if you want to avoid paying for every click.

At least 50,000 page views per month

On your website, excluding in-app and login data. This makes the effort of personalization (lots of content variations for lots of audiences) worthwhile vs. making generalized, quicker, “global” improvements.

At least 100 conversions per week

As with traffic, we recommend you already have a significant number of conversions from your website. This indicates your website is already performing and can be optimized vs. getting the foundations of your marketing funnel in place.

Previous experience with A/B testing for web

This is helpful to understand what’s more easily implemented and what’s likely to drive bigger wins with less effort. You might also have documentation from previous A/B tests which is useful for designing personalized experiments.

Results we’ve seen

Amongst B2B SaaS companies implementing real-time web personalization

  • Doubled signups
  • Tested 105 ideas with 4.5 billion page combinations in 3 months
  • 79% increase in customer referrals within 3 months
  • 55% increase in conversions with an increase in Average Customer Value (ACV)

More "Spotted" playbooks like this?

Subscribe to The Crow's Nest newsletter from Hull for best practices, ideas & resources for customer data management.

What you should do now

  1. If you'd like to setup Hull - to seamlessly sync all your tools, teams & data (like we did for all these companies), then talk to sales to setup a free data flow consultation.

  2. If you'd like to learn our best practice for customer data integration, read our free Guide to Getting Started with Customer Data Integration.

  3. If you enjoyed this article, perhaps your team will too? Why not share it with the links below.

Ed fry
Ed Fry

'Ed of Growth at Hull, working on all things content, acquisition & conversion. Conference speaker, flight hacker, prev. employee #1 at inbound.org (acq. HubSpot).

If you've questions or ideas, I'd love to geek out together on Twitter or LinkedIn. 👇