Conversational Advertising: The Definitive Guide [2019]

by | Feb 25, 2021 | Uncategorized | 0 comments

This is the most comprehensive guide you’ll find on conversational advertising.

If you’re currently running some combination of programmatic display ads, social campaigns, and static landing pages – this guide will teach you everything you need to know about how conversational advertising works, how to put it into action, and most importantly, answer one simple question:

“How can I drive higher engagement from my digital campaigns?”

Full disclosure: this post also tells the story of our conversational ad platform. So yes, a guide to the benefits of conversational advertising is self-serving.

We’ve created a table of contents in case you want to jump around and go directly to a section that interests you most.

Let’s get started…

Part 1: How Did We Get Here?
Part 2: What is Conversational Advertising?
Part 3: Why Does it Work?
Part 4: Conversational Display Ads
Part 5: Conversational Social Campaigns
Part 6: Conversational Landing Pages
Part 7: Measurement & Analytics
Part 8: Future Formats

Part 1: How Did We Get Here?

Here’s a simple question:

“What would online advertising look like if the customer experience came first?”

Banner ad formats

It’s unlikely anyone would come up with a banner ad as the answer.

People love talking about the best TV spots on the Super Bowl. You might even remember clever print or radio ads.

What about digital? Do you remember the last time you saw – or engaged with – a compelling online ad? If the answer is no, you’re not alone.

Check out some of these stats:

“Most online ads these days are insulting to my intelligence.”
“Most ads I see online don’t look polished or professional.”
“I don’t notice online ads anymore, even if I don’t block them.”

Source: HubSpot, “Why People Block Ads (And What It Means for Marketers and Advertisers”, 2016/20181

Of people who have clicked on an ad, 34% said it was a mistake and 15% accused advertisers of tricking them into clicking. Only a paltry 7% said it was because the ad was compelling.1

The Customer Experience Isn’t First

25 years after the first banner, we have…pretty much the same banners.

While the media side of the business has evolved dramatically and is incredibly complex, the creative formats and user experiences we started with haven’t really changed.

Perhaps this is an over-simplification. And some of these formats undoubtedly work for some brands. But the numbers around consumer engagement don’t lie.

Here are three big why we believe this is so:

The customer experience is an afterthought.
We’re stuck with rudimentary ad formats that literally have not changed in decades. Hence banner blindness.2
It’s reductive. The industry has tried to automate and reduce a complex problem down to an image, headline, body copy, and button that will ‘just work’ everywhere.

The really crazy thing? This isn’t just bad for consumers, it’s terrible for advertisers too! In the examples above, that tiny Lincoln logo can’t be doing much for the brand. And the Four Seasons banner doesn’t say ‘Four Seasons’, to me at least. Yet look at the current and projected ad spend:3

There must be better ways to do this. Here’s one (admittedly biased) possibility…

Part 2: What is Conversational Advertising?

For us, customer-first online advertising should:

Provide value, or at the very least, utility, for a consumer.
Not disrupt a consumer’s browsing experience.
Be respectful, and ideally deserving, of a consumer’s time and attention.

Which Brings Us To Conversations

Conversations are how we naturally engage with each other. Written or spoken, they are our primary form of communication.

Plus, conversations (at least those that we willingly engage in) tick the boxes of what we said online advertising should be:

Provide value or utility.
Not be disruptive.
Respectful of our time and attention.

It seems like a decent starting point.

What Exactly is Conversational Advertising?

Conversational advertising is the use of automated, natural language conversations between brands and consumers using conversational assistants (chatbots) to drive awareness, purchase intent/consideration, and conversions.

Here’s How it Works…

Automated conversations (those that are powered by machine learning and natural language processing) are made up of ‘intents,’ which are the basic building blocks of a conversation.

Every intent has two parts:

An utterance, or what a user says.
A response, or how the chatbot answers the utterance.

Here’s a simple example of what an intent for an automotive conversational assistant might look like if a user asks about towing…

Chat Utterance
Every Intent has an utterance…
Chat Response
…and a response

While user utterances are typically text or voice-based, responses can be delivered back to the user as any combination of images, text, cards, and buttons – or custom-built, specialized response formats.

AdChat Response - Card
Cards combine images, headlines, text and an optional button.
AdChat Response - Text & Emoji
The full range of emoji and text for utterances and responses.
AdChat Response - Image & Video
Inline media, including images, animated GIFs, and video.
AdChat Response - Buttons
Buttons that either drive to intents in the chatbot or to external URLs.

A conversational assistant, at its most basic, is a bunch of intents that together, help consumers get relevant information about a brand/product/service in a natural, conversational way.

Of course, things can get more involved – for example, if you’re collecting leads, allowing users to upload images, or pushing consumers down a specific path/funnel. 

Acknowledge What Isn’t Understood

Many conversational assistants annoy consumers when they can’t understand the user’s utterance. Often, they fail even to acknowledge what the user said and that it was not recognized.

High-quality conversational interfaces use a combination of machine learning (ML) and natural language processing (NLP) to understand the context of an utterance so that a useful and relevant response can be provided.

But of course, there are times where users either don’t make sense, or the conversational assistant simply doesn’t understand what is being said.

Having thoughtful ‘fallback’ responses to these situations is crucial for building a good user experience.

Here’s an example:

Chat Fallback Response 1
The chatbot asks a simple question…
Chat Fallback Response 2
But the utterance isn’t what the chatbot was looking for.
Chat Fallback Response 3
The chatbot is able to ‘gracefully’ provide a fallback response.

Now that we have a basic idea behind how automated conversations work, let’s explore why these experiences can be engaging for consumers.

Part 3: Why Does Conversational Advertising Work?

Anyone who has used a Messenger chatbot knows just how high the engagement can be.

84% open rate of Messenger messages
Open rate of messages sent using Messenger4
55% of consumers prefer talking to a business through a chatbot
Consumers who prefer to talk to businesses using a chatbot4

Why Chat-Based Experiences Are So Engaging

Here’s our take on why consumers engage with conversational assistants:

Interacting with a conversational assistant is a natural extension of texting and messaging. It’s familiar and intuitive.
Consumers can ‘self-select’ how to engage – whether that’s through clicking buttons, free-form text input, or voice.
Consumers can ‘self-select’ the content they’re interested in. Compare that to a static landing page – what are the odds that the info on that page is relevant and top-of-mind for every visitor?

It’s the third point that really sets conversational advertising apart. It’s no longer about a ‘one size fits all’ ad or landing page. Chatbots can deliver personalized, highly relevant content in a naturally engaging way.

Now let’s take a look at the different ways that conversational advertising can be brought to life.

Part 4: Conversational Display Ads

“It’s like a chatbot and a display ad got together and had a baby.”

On the surface, putting a conversational assistant into an ad would seem like a (relatively) straightforward integration. But it’s not.

Creating Conversational Display Ad Formats is Technically Challenging

There are a lot of technical hurdles to overcome when developing a new display ad format. Here are just a few:

Ensuring the ad format falls within commonly accepted file size limits.
Limiting the number of server calls/HTTP requests; as you can imagine, there is a lot of back and forth between the ad and server with a chatbot.
Making sure the format is accepted and works consistently across DSPs, ad servers, and SSPs.
Ensuring speed/quick responses to user utterances within the ad.
Ensuring ads can be deployed programmatically/as part of a programmatic buy.

And…There are Some User Experience Challenges

On top of the technical considerations, there are also significant user experience challenges.

Chief among them is creating a UX/UI that works seamlessly across a range of ad sizes/placements.

In theory, conversational display ads can work at any size. However, we’ve found that:

300×250 delivers a ‘decent’ user experience, but it’s the absolute smallest size that makes sense.
336×280 is better – it ensures some extra room for the various response formats.
300×600 and larger is best. Responses can all fit into the frame at once and are easily read.
300x250 Conversational Display Ad
300×250 is the ‘minimum viable size’ for a decent user experience.
336x280 Conversational Display Ad
336×280 is a better size as it allows room for all UI elements.
300x600 Conversational Display Ad
300×600 and larger (including expandables) is best.

Expandables and other large formats also work well to initiate the chat experience in an ad.

With that, let’s take a look at bringing conversational advertising to the ‘walled gardens’ of social apps to get more out of social campaigns.

Part 5: Conversational Social Campaigns

Options are pretty limited when it comes to ad formats and generating unique experiences with social ads.

Although the basic social formats are arguably better than standard display ads, they still don’t offer much in the way of engagement over and above a click, like, or comment.

Instagram, Facebook, Twitter, and LinkedIn In-App Conversational Advertising

What if your social ads could trigger an in-app conversational experience on Twitter, Facebook, Instagram, or LinkedIn?

Social Icons

A chat-based experience that runs in-app feels like a natural extension of the social platform and is a less disruptive browsing experience.

There are a couple of different ways these conversational experiences get deployed in-app. 

For Instagram, Facebook, and LinkedIn, the conversational experience gets triggered by an ad and loads in the in-app browser. Depending on the campaign goals, the entire customer interaction can remain in the chat, or consumers can jump over the advertiser’s site at appropriate times in the conversation.

For Twitter, the integration can either be through the in-app browser, or a native integration.

What’s the difference? The native integration uses Twitter’s UI, so everything looks and feels completely seamless. And users can engage with the chat session through an ad, or by sending a DM.

Twitter DM API Integration
Twitter native integration uses Twitter UI (via the DM API).
Twitter In-App Browser Integration
Twitter in-app browser integration uses customizable AdChat UI.

Replace Static Landing Pages With Conversations

Most social ads direct users to a static landing page on click. And a lot of times these landing pages are a ‘wall of text’ that the advertiser hopes the consumer is interested in reading.

Therein lies the problem. It’s just not realistic to expect that every consumer is going to get the information they want or need from the same static landing page.

By replacing these static pages with in-app conversations, the ‘one size fits all’ nature of the experience is gone – and replaced with something that is engaging, highly personalized, and non-disruptive.

Let’s take a look at one more way to use conversational advertising – as a destination for other marketing activities.

Part 6: Conversational Landing Pages

Most marketers do more than run display ads and social campaigns.

But what if all channels – programmatic buys, Instagram ads, emails, direct mail, and print ads – had a consistent, relevant destination that engaged consumers in a conversation? 

Build Once, Deploy Everywhere

The benefit of building a chatbot on a conversational ad platform is that it can be created once, and deployed in multiple environments.

It can be used as a stand-alone experience on a landing page, or be integrated into an existing page or pages throughout a website

A conversational landing page experience works in exactly the same way as display ads and social campaigns.

An added benefit is the ability to target different parts of the conversation depending on the channel consumers are coming from.

For example, an email announcing a sale on multiple products might drive recipients to a product recommendation quiz within the landing page chatbot. Social ads might drive to another section of the chatbot entirely.

The point is, each channel or even different ads or campaigns in a channel can all drive to a different conversation within the same bot.

For one client (Troy-Bilt agency Marcus Thomas), we’ve built a single conversational chatbot that runs in Messenger, in programmatic display ads, and as a landing page. In each channel, the ‘starting point’ of the conversation is different.

Here’s how it’s used on a landing page for Troy-Bilt’s various voice and chat deployments:

Troy-Bilt Conversational Landing Page

Part 7: Measurement & Analytics

None of this much good without a solid way to measure engagement and conversations.

At the very least, analytics should measure:

Active users over time (monthly, weekly, daily).
Sessions over time.
Average session time per user.
Average messages sent per user.
Exits, with URL and count.
Conversation Flow.

Let’s take a closer look at the last two on that list. Most of the others are self-explanatory and are familiar territory with products like Google Analytics.

Conversation Flow and Transcripts are related. They both track every interaction or message sent to the chatbot.

Measuring Conversation Flow in Conversational Ads

Conversation Flow measures interaction on an aggregate level, breaking down the most popular messages sent to the chatbot. In the diagram below, we can explore various paths that users have taken through the experience before either leaving the chat experience via clicking on an exit URL or by leaving the session.

Session Flow

Aside from tracking how users successfully engage with the conversation, session flow is also useful for advertisers and agencies because it lets them identify any ‘missing gaps’ in the conversation, as represented by the red bubbles in the above diagram.

In this automotive example, some consumers are asking about technology, but the chatbot initially did not know how to respond. By identifying this need, tech-related content was seamlessly added to the experience to fill in this missing area.

Transcripts in Conversational Advertising

Transcripts break down session flow by individual users or sessions. Every time someone engages with a bot, a unique session ID is generated. This is a long alpha-numeric string. There’s nothing personally identifiable about it.

Transcripts allow a marketer or agency to see how individual users move through the experience. This can reflect areas where people abandon a funnel and what they say to the chatbot at that point in the conversation.

Content in the bot can then be added or edited to fix problem areas. Below is an example transcript showing a consumer that sent 3 messages to the chatbot (green), and the intents that were triggered as a response (grey).

Chat transcript

Part 8: Future Formats

While conversational advertising is still new, there’s no question that the future lies with voice assistants.

It remains to be seen how Google Assistant and Amazon Alexa will integrate advertising into their ecosystems, but it will undoubtedly happen.

Conversational Assistants

Here are some key stats on the penetration of voice assistants. With a market this big, advertising can’t be far behind:

Over one billion voice-enabled devices were purchased in 2018.5
It’s estimated that 55% or 70 million U.S. households will have a smart device like Amazon Echo, Google Home or Sonos One by the year 2022.6
An estimated $40 billion in purchases will be made through voice assistants in 2022.7


Conversational advertising is one of the best ways to generate a positive return on ad spend (ROAS), whether you’re building a brand, launching a product or service, generating leads, or remarketing. 

It’s a highly engaging format that allows consumers to self-select the content they’re interested in, increasing the likelihood that they will engage and walk away from the experience better-informed.

Interested in learning more? Book a demo to see how AdChat’s conversational ad platform can bring it all to life.