We’ve been doing conversational marketing since before it even existed.
We learned it from the ground-up, starting on the phones 30 years ago and transitioning into live chat when the first tools landed.
Now, we are at the forefront of AI technology in a conversational world, pushing an average of 9x more conversions for our clients and users.
We have learned a lot along the way, and we continue to learn everyday.
Here’s what you should know about conversational marketing in 2022.
The undeniable truth: Your website needs conversational marketing.
Over 80% of B2B companies are using conversational marketing (also known as ‘chat’) on their website.
Why? Because times have changed.
Website visitors and customers expect an immediate response.
They don’t want to fill out a form and wait, and they certainly don’t want to pick up the phone.
We live in the age of instant gratification. WhatsApp, Messenger, Slack, and Teams.
We communicate in real-time and in a conversational way.
Chat generates incredible results.
There’s no doubt that chat works, and works well.
In fact, it’s one of the best possible ways to increase your website’s conversion rate compared to other techniques like pop-ups, and UX changes.
- Customers who use live chat spend up to 60% more per purchase than those who don’t. (Source)
- Live chat improves conversion rates by 3.84% on average, using the traditional methods available today. (Source)
- 44% of online consumers say that "having questions answered by a live person while in the middle of an online purchase is one of the most important features a website can offer”. (Source)
- 38% of consumers are more likely to buy from a company if they offer live chat support. (Source)
But, chat is hard to do right.
Most companies buy a chat solution and only scratch the surface of what they’re capable of.
Why? Because chat is hard to do right - especially at scale.
Many marketers have asked one or more of these questions:
- How do you know when to use a chatbot vs a live resource?
- How do you know which visitors could end up in a sale if provided help from a sales resource vs which visitors would have bought anyway?
- How does your BDR team avoid wasting time chatting with the wrong visitors?
- What pages should have chat implemented?
- What customer journeys should have chat implemented?
Chatbots help, but they’re far from perfect
One of the key tactics used by companies today is to use automated chatbots to help their customers. Make no mistake, automated chatbots play an essential role in a well-rounded chat strategy - but they don’t perform as well as a real person when it comes to converting visitors into sales or leads. 38% of businesses say their users find their scripted responses most frustrating. (Source).
In fact, live sales agents are 2.5x times more likely to convert a visitor into revenue than an automated chatbot.
Chatbots offer highly scripted responses that don’t cater to the website visitor’s specific needs. You feel like you’re being funneled through a series of hurdles to get what you actually need, and leave you feeling frustrated. If you were on the cusp of buying, you can imagine how this might reduce the chances of that happening considerably.
Our experience shows that up to 65% of people who purchase from a live sales resource, do so because they have a specific use case or need addressed by a salesperson that can’t be served by FAQ’s or a bot experience.
However, automated chatbots are great at offering support and nurturing visitors. For example, if a customer has a frequently asked question, chatbots can send them a helpful article or even direct them to a live sales agent.
That said, if a visitor is on the cusp of buying, but needs to have something specific addressed or needs to be reassured, you’re far better off connecting them with a live agent immediately, rather than being stuck in “chatbot jail”. Here’s a more in-depth look at the world of chatbots today.
“Conversational AI” isn’t there yet, either.
One hot technology in conversational marketing is ‘Conversational AI’, which aims to recreate a human-like conversation.
These models use various techniques to accomplish this - from language and semantic processing to complex conversation “trees” that take into account common “branches” of conversation to reach a desired result.
The problem is - these models take a significant amount of time and training to be useful, and in a sales environment are a long way off recreating a human experience.
Simulating a real human is no simple task. They require a lot of upfront training just to launch and then ongoing training to improve. So, they slowly learn with each conversation what kind of questions your visitors ask and how to get them from point A to point B. This usually takes thousands, if not tens of thousands of interactions to get any material learning.
That said, conversational AI should not be ignored. When used correctly, it can be very powerful in conjunction with other conversational marketing tools and techniques. In fact, a large cohort of site visitors can be managed by conversational AI. It has a significant role to play - but again, it’s currently best used on visitors who do not need the human-touch.
Page-based rules or complex journey mapping don’t work well.
So, if you can’t use chatbots across the board to optimize results - what do you do?
Most of the chat tools you use can be programmed to send invites (chat requests) to visitors based on various rules and criteria. If these criteria are met, a chat “playbook” (template) is triggered - such as connecting visitors with a live agent. For example:
These allow you to send chat invites on certain pages only, such as the pricing page, product page, or demo page. The thinking is that visitors must be interested in your product or service if they’re viewing these pages, so offering chat can help get them across the line in those key moments.
However, there are two critical problems with this approach:
- Coverage. This refers to how many of your visitors are sent a chat invitation across your website. Typically, a page-based approach will net you a small percentage of the potential visitors who could be converted into revenue. For example, if you only present chat on the pricing page and contact page, only a handful of your total visitors will see or engage with those chat invitations, and you will miss all the potential conversion opportunities that are on other pages.
- Conversion. If you managed to get 100% coverage by using chatbots at every possible instance, your conversions would suffer because your customers will be frustrated with the experience (as discussed above). Similarly, if you hired enough salespeople to be present on every page of the site, they would end up being frustrated as not all visitors want or need their assistance (not to mention it would be incredibly wasteful use of your valuable sales resources).
User Journey Mapping
This is when you program the chat tool to send invites based on the visitor reaching a more complex series of criteria than just page-based rules. For example:
- Visitor spends more than X amount of time on site
- Visitor visits a minimum of X pages on the site
- Visitor looks at X key content blog
Again, there are two problems with this approach. The first is the same simplistic hypothesis of a page-based approach. A human (usually a marketer) is making an informed guess as to what constitutes buying intent. However, these steps and criteria in the buyer journey are often arbitrary - especially when users navigate websites in incredibly complex ways - often with multiple tabs open comparing different offers at different stages of the buyer journey.
The second approach is that it’s incredibly time-consuming and complicated to set up all of these user journeys, keep them updated, measure the success of them, and manage them. The more complex the site and the more traffic you have, the more this becomes impossible.
You’ll also find that the user journey mapping approach results in low coverage, similar to page-based rules.
So, most marketers resort to asking for the visitor’s email address.
If you’re not maximizing coverage, conversions, or revenue from your chat strategy - you might as well get the visitor’s email address, right? That way, you can market to them using automated email tools and can satisfy your sales team's desire to qualify everyone who is sent to them.
But, there’s a problem.
Up to 80% of your visitors will abandon chat if they’re asked for their email address too soon.
Why would someone give up their personal information, knowing they’ll probably be marketed to, if you haven’t first offered something of value to them? This is the problem with asking for an email address too soon.
When you walk into a store, does the staff member immediately ask you for your phone number? No, they read the room, determine if you need help, and make an offer of value. The same should be true of an online experience.
So, asking for an email address isn’t going to get you optimized results either.
The underlying problems in chat today.
Most of these complex toolsets, hypotheses, and tactics only scratch the surface of what’s challenging in chat. They partially solve some of the symptoms of two underlying problems that haven’t been addressed:
98% of website visitors are anonymous
According to Marketo, up to 98% of your traffic is anonymous - meaning you have no idea who your visitors are. Imagine customers walking into a store, but they’re shrouded in cloth and you can’t make out who they might be. That means you’re treating visitors with a generic, “one-size-fits-all” approach. If you did know who the visitors are, you would set up specific playbooks just for them.
Visitors have real-time expectations
The other key problem is that visitors expect to be helped in real-time. That means right now, in the moment. However, you can’t possibly help every single visitor on your site right now - especially if you have thousands or tens of thousands of visitors. You simply don’t have enough sales resources to connect with every visitor (nor would they want to). And as we learned, chatbots and conversational AI aren’t the answer.
What about identifying visitors using a traditional ABM tool?
To address the “anonymous” problem, companies make use of identity intelligence tools such as Clearbit, Zoominfo and 6Sense. These are casually known as ABM tools.
These tools reference databases of information in order to determine who a visitor is on your website. For example:
- Visitor’s IP address is known to belong to XYZ company
- This company has XXX number of employees
- This company is known to be in XYZ region, and ABC country
- Visitor is part of XYZ target industry
- Visitor’s IP address has been seen before with this email address: email@example.com
- Visitor came from XYZ paid search campaign
- This visitor fits our ICP
With this information in hand, marketers and sales teams can take action. They might recognize that this visitor is from one of their target accounts, so they can initiate a chat engagement with them or set up retargeting ads after they’ve been on the website.
However, there are four key problems with the ABM tool approach:
- The database referenced by your ABM tool might be out of date. That company, person, or IP address may be incorrect or irrelevant now. This is especially true in the post COVID work from home world, where IP address information is not necessarily attached to a company anymore.
- You still don’t necessarily know exactly who is on the website. While your ABM tool is shouting “look, this person is from Coca Cola!” you probably won’t know who the actual visitor is and it could be difficult, if not impossible, to track down the actual person..
- Even if you know what company they are from and who they are, you still don’t know what their buying intent is and what resources to engage them with. You can make some assumptions based on a hypothesis, but that is just scratching the surface of really understanding the visitor’s buying intent.
- Even if you know who the visitor is and what their intent is, you aren’t necessarily going to engage them in real-time, when it matters. That’s because you either don’t have enough sales agents, the visitor was sent a chat invite but declined, or the visitor got stuck in “bot jail” instead of being directed to a live agent.
What about following up with “hot” visitors in my CRM or lead scoring tool?
Knowing that Jane from Coca Cola was on your website 6 hours ago isn’t necessarily a good reason to send her a follow-up. She may have expressed no purchase intent, or she might be in a very early stage of her buyer journey, or she was merely passing through.
Most importantly though, you missed striking while the iron is hot.
CRMs come bundled with helpful tools like lead scoring to help you understand the warmth of a lead or visitor, but this is only useful after the visitor has left your website and the CRM contact information has updated.
Even after that, it’s only useful once your sales team takes action on it. Here’s a larger breakdown on the drawbacks of lead scoring.
Remember, visitors have real-time expectations, so although following up with target accounts is a valuable tactic, you should be prioritizing those visitors in real-time chat as much as possible.
What about engaging visitors in real-time using Drift, LivePerson, or other chat tools?
To solve the ‘real-time expectations’ problem, you could (and should) make use of conversational marketing platforms such as Drift, LivePerson, Intercom, etc. You might call these “chat tools” but today they offer functionalities that go well beyond a basic chat tool, and have become essential platforms for engaging customers on your website.
This article has largely discussed the merits of using chat tools on your website, and although they do offer the ability to engage with visitors in real-time (similar to a sales person engaging with a customer in a store), they struggle most when trying to engage anonymous visitors.
Why? Because although you have the ability to engage any and all visitors, you have no idea which exact visitors to target.
You can, of course, set up conversational playbooks that trigger based on data from your intelligence tools (such as known accounts, industries, etc), but typically this data only helps to identify on average 30% of your traffic (and even then, you’re working with surface-level data).
Here’s a visual of what a very mature conversational marketing process looks like, using all of today’s relevant tactics and technologies:
This visual represents the culmination of the tools and techniques used by the most advanced companies today - making use of page-based playbooks, ABM tools, qualifying playbooks, and chatbots. These tools are all useful in their own right, but even the most sophisticated users are not fully optimizing their capabilities because most visitors are anonymous, they expect real time engagement, and until now it was impossible to effectively determine every visitor’s buying intent - so marketers default to asking those visitors for the email address.
Enter the Conversational Marketing Solution: Buyer Intent + Artificial Intelligence
As discussed, the vast majority of challenges in the conversational marketing space stem from two underlying causes: most of your visitors are anonymous, and they expect real-time engagement. We can’t solve those problems by throwing chatbots at every visitor, or using premature conversational AI, and we can only identify a small percentage of those visitors using ABM tools (and even then, that doesn’t guarantee that you have actionable information on the visitor).
So, how do we truly solve these problems?
It starts with buyer intent.
Solving for Anonymous Visitors: Buyer Intent Data
Remember, over 70% of typical website traffic is completely anonymous (unknown to your ABM or martech tools).
If you have 100,000 visitors coming to your website a month, that’s 70,000 anonymous visitors per month that you’re hoping will fall nicely through the cracks of your website to convert by themselves.
The typical conversion rate for that experience is 1-2%.
Put simply, you’re leaving far too much money on the table.
But, imagine if you could suddenly reveal the intent of all of those visitors. You could see who is ready to buy, who is on the fence, and who you can serve with content or an automated experience.
This is known as “buyer intent”, and to determine it accurately for each visitor, you need the right data.
What is Buyer Intent and What are the Alternatives?
Put simply, buyer intent is the propensity of any web visitor to buy or convert on your website.
For example, someone with a high buyer intent is on the cusp of purchasing or signing up for your solution.
Someone with low buyer intent is still early in their evaluation process for different solutions/products, or they could be searching for customer support.
Buyer intent as a metric is determined by the quality of ‘buyer intent data’ (the variables that are used to inform the result), and then the algorithm or model used to calculate that result.
Most of the buyer intent tools on the market today are based on the research the prospect is doing, and is gathered from 2nd or 3rd party data vendors. For example, companies like TechTarget and Bombora will tell you when a prospect is researching solutions similar to yours.
Other vendors like 6sense also look at what keywords that prospects are searching to solve their problems and can tell you if a visitor used these keywords to land on a relevant website, if they recently came to your website or a similar site, and other measures.
These buyer intent tools are good but there are two problems associated with them.
One is that typical buyer intent tools are aggregated by a third party so you are purchasing data and information that is also likely being sold to a competitor and may be out of date. Often you will access the data past the point where it is most useful (and all your competitors have it as well.)
The second problem is that determining the buyer's intent is most often a human-based hypothesis. Somebody, somewhere, decides that this particular range of keywords must express a high buyer intent, and these on-page behaviors do too. Add them together, and you must have a ready-to-buy customer, right?
Buyer intent is far more complex than just a few surface-level measures. It requires looking at thousands of variables, calculating them in real-time, then surfacing that data to your engagement tools so you can take action.
Machine Learning and AI Is The Only Way
No human can possibly crunch those kinds of numbers, and if they could they certainly couldn’t do it on the fly before the visitor leaves the site.
Neither can simplistic algorithms that use a static equation to calculate a result. These algorithms are again based on a hypothesized idea, derived by a human, which add, subtract, multiply, divide, and perform other functions that can't possibly account for the nuances of visitor behavior.
That leaves machine-learning and artificial intelligence.
Using a model trained on millions, if not billions of data points, a machine-learning model can work backwards to determine which visitors have buyer intent.
The model can tell the buyer intent mode what a successful sales journey looks like and what a failed buyer journey looks like, then feed the model vast quantities of data to continuously learn and determine what behavior typically ends up in a sale, lead, or support request.
Using this approach, machine-learning models can simultaneously look at historical data and real-time, on-page behavior of visitors to predict their buyer intent with incredible accuracy.
The Perfect Conversational Marketing Strategy
Put this all together, and you have the recipe for a perfect conversational marketing strategy that generates revenue from both the known and anonymous visitors (who make up the majority of your website traffic) on your website.
Anonymous Buyer Marketing (The New ABM in Town):
The concept of assigning buyer intent scores to anonymous visitors using AI, then converting them into revenue through chat is completely new in the marketing technology landscape. For the purposes of making this clear, we have coined the term “Anonymous Buyer Marketing, or the new ABM.
Here’s how it works:
- As visitors navigate the site, they are assigned a “buyer intent score” by the AI model
- Those visitors are segmented into high, medium, and low buyer intent (including completely anonymous visitors and ABM-identified visitors)
- Based on those segments, your chat tool can deploy a playbook designed to align with the visitor’s needs
- High buyer intent visitors should be sent directly to your live agents in real-time for the best chance of converting
- Medium buyer intent visitors should be sent to a combination of chatbots and live agents in real-time for nurturing into leads and sales
- Low buyer intent visitors should be sent to fully automated chatbots in real-time that offer helpful content, FAQs, and conversational AI
Supercharge your ABM, CRM, and MAP tools
The same logic for anonymous visitors can be applied to your known visitors:
ABM Known Visitors
- Your ABM tools can now come alive, adding further segmentation to their demographic and technographic data.
- Known accounts that have high buyer intent can be prioritized above everything else, sent to Account Executives or similar for conversion through chat
- Known accounts with medium buyer intent can be sent to live BDR agents for nurturing
- Known accounts with low buyer intent can be nurtured through chatbots and content
Known Visitors Missed Through Chat
If the buyer intent AI + real-time engagement strategy missed visitors (because they didn’t accept chat or exited the site too quickly) then you can still take advantage of buyer intent scores in your CRM and MAP tools:
- Contact information in your CRM can be updated with the visitor’s buyer intent score
- Contacts who were on your site and had high buyer intent can be prioritized with outbound motions (sales outreach, retargeting ads, automated email marketing, etc)
A World-First Buyer Intent Technology To 9X Conversational Marketing: Lift AI
Lift AI is the only technology on the market that uses a machine-learning model to determine the buyer intent of visitors both known and anonymous, in real-time.
Lift AI’s proprietary machine-learning model was trained using billions of data points and millions of live sales interactions, gathered over 15 years.
That data was codified into the machine-learning model and pre-trained to determine a visitor’s buyer intent with pinpoint accuracy.
Lift AI works in real-time to assign every single visitor a “buyer intent score” which is then fed into chat tools like Drift to deploy customized playbooks depending on the score.
No other tool can currently do this, because Lift AI’s data was gathered in-house (not purchased from second or third party data vendors which often carry outdated and incorrect data.
Based on Lift AI’s findings so far, a typical website has the following breakdown of visitors:
- 9% have high buyer intent (how we know this)
- 17% have medium buyer intent
- 74% have low buyer intent
So, as you read this, a significant amount of your traffic is on your website expressing an intent to buy or become a lead - and you have no idea who they are or how to engage them.
Consider how many visitors are coming to your website each month, and some quick math will reveal the extent of your opportunity.
In fact, on average, companies using Lift AI with a high-performance conversational marketing platform like Drift get a 9x increase in conversations to pipeline.
It’s impossible to get these results without Lift AI.
Here’s a typical overview of what a website looks like with a standard implementation of Drift vs an implementation including Lift AI:
Without Lift AI:
- Conversations: 228,334
- Opportunities: 948
- Pipeline Conversion: 0.4%
With Lift AI:
- Conversations: 19,601
- Opportunities: 727
- Pipeline Conversion: 3.7% (9x increase!)
Put this all together, and the below visual represents what a revised conversational marketing pipeline looks like with Lift AI, where everything under the large purple arrow is now “intent-based” and leverages all the other existing tools and techniques to create a transformational result:
If you want to dig deeper into our 9x conversion numbers, we conducted a full study and data analysis here.
Results are not only game-changing, but they can be achieved quickly too. PointClickCare was able to generate 4x more chat conversions in just 90 days.
Similarly, Formstack was able to increase their pipeline from chat by 88% (also in 90 days).
These results are unheard of in the digital marketing space.
Some companies spend tens of thousands doing pop-ups, UX/UI changes, and more to squeeze another 1% of conversions out of their website.
So why isn’t every company using this conversational marketing strategy to get 9x results?
Because the vast majority don’t realize that this opportunity exists - largely because artificial intelligence has evolved so quickly - but also because there’s only one tool in the world that can execute this strategy.
So, instead of flying blind and spending resources on outdated conversational marketing strategies, you can start implementing an intent-led strategy right now using Lift AI.
Get started - for free. We stand by our numbers and promises.
Lift AI is simple and lightweight to install - it’s a short snippet of JS code that is pasted into the </header> section of your website, and it immediately starts assigning visitors their buyer intent scores.
With Lift AI, you will also receive a free Revenue Opportunity Assessment after 30 days where you can see how much revenue you are missing, and document a plan to capture the missed opportunity