Manufacturers and other organizations whose focus is selling business-to-business, need to ramp up their web presence to meet the demands of educated buyers. Your organization needs to move from improving your web presence to increasing mindshare and attracting new customers by creating immersive, personalized online experiences. Whether you build tablet presses, CNC lathes, medical devices, or power tools – visitors to your website are in need of personalized content to continue their journey, with your company top of mind.

Based on Innover’s industry experience, the following framework can enable your company to shift from an academic web presentation to an immersive experience for every visitor to your site. In other words, moving from presenting an informational website, to directly impacting top-line growth through Experience Optimization.

There are three core components of Experience Optimization:

• Data Strategy
• Content Strategy
• Personalization Strategy

Data Strategy

Your company has a lot of data. Transactional, financial, operational, third-party, behavioral, and web analytics data.

Your data strategy could include the following considerations below – not an exhaustive list by any means, but a good start. We are intentionally not including data activation as a strategy, as we’ll cover that under Personalization Strategy.

Data Strategy Considerations:

• Do you have a record of all your company’s data sources and where the data originates?
• Where does your company store its data, e.g., data lake, data warehouse, customer data platform, cloud storage?
• What are the technologies and languages used to extract, load, and transform your data (ETL or ELT depending on strategy)?
• Is data engineering/DevOps a need?
• Is the data categorized in a way that easily identifies which data is structured, semi-structured, and unstructured?
• Is there a codified data governance framework in place yet?
• Is data science as a service or through a cloud platform a need?
• Is stitching data into a unified profile for segmentation and audience building a need?

Content Strategy

We’ll discuss content strategy in the context of customer-facing content. Your company may or may not have a significant number of digital assets (jpg, tiff, psd, mp4, png); however, if you combine your digital assets with your technical documentation, product documentation, safety sheets, user guides, and PDF catalogs, all of a sudden you have a content challenge.

Your content strategy could include the following considerations below – not an exhaustive list by any means, but a good start. We are intentionally not including content activation, as we’ll cover that under Personalization Strategy.

Content Strategy Considerations:

• Do you have a library of all of the content your company shares with your customers online?
• Where is online content stored, e.g., CMS, DAM, cloud storage, internal servers, on employee laptops, thumb drives?
• Is there a codified content governance framework in place – if not, should there be one?
• Are there permission-based rolls in place to disposition content access, editing of content, downloading content, and adding content?
• If the content is indeed consolidated in a repository – is there a taxonomy in place that can be easily traversed and searched by all users in your organization?
• Does anyone outside of your organization have access to your content, if not, should they?
• Does your organization have an efficient way to share content among creative teams for collaboration?

Personalization Strategy

This is where the rubber meets the road for effective execution of Experience Optimization. You can have an experience without personalization, but you cannot have personalization without experience. Companies can create experiences, but can they prescribe these experiences in a way that feels authentic? If not, the experience your visitor receives might not be a “good” experience. It might elicit the opposite effect. From Innover’s experience, your Personalization Strategy is a component of your Experience Optimization Strategy, and it should be viewed as separate but equal to your Data & Content Strategy.

We will not be proposing questions below for your organization to reflect upon – instead, we are providing declarative, unambiguous requirements for true one-to-one personalization that your organization needs to employ to deliver on Experience Optimization.

• Your organization must have a data strategy and a content strategy in place before you can consider a personalization strategy. Content without context is meaningless. Your data is your context. If someone is found walking without a coat in the desert would you give them a blanket or a fan? You cannot answer that question without context – without data. 2 pm in the desert might be oppressively hot, whereas 2 am in the desert might be dangerously cold.
• Your organization has to employ some type of personalization technology. Now that could very well be a recommender system that your IT team has coded in Python and leveraged via API from your own experience platform. I mean if you simply apply a linear combination of Matrix Factorisation (a.k.a. SVD) and Restricted Boltzmann Machines (RBM) you should be able to achieve a root mean squared error (RMSE) of 0.8 to 0.9. What?!? Just invest in the out-of-the-box technology that leverages prebuilt AI-applied Machine Learning, i.e., Random Forest Algorithm and Multi-Armed Bandit.
• Your organization should employ a very competent agency/SI partner to implement the technology. A poor implementation means trouble from the start, and it could mean never actually launching the technology. It’s like hiring an employee who refused to come to work every day for a year and the work piled up on their desk.
• This partner should be able to integrate your personalization technology with your analytics measurement and optimization technology. Perhaps that is integrating your Salesforce CRM with Salesforce Interaction Studio or Adobe Analytics with Adobe Target or Google Analytics with Google Optimize360 or any permutation of the above technologies if the partner has the experience.
• Your organization should work towards consolidating all of your digital assets and technical documentation into one central repository, preferably an enterprise Digital Asset Management solution.
• Utilize a DAM that employs Machine Learning to give back time to your stakeholders. For example, Adobe’s DAM utilizes a Deep Learning image recognition algorithm to help find unlabeled assets and automatically add metadata to those assets. The ROI is tremendous – 366% within 3-years.
• Your DAM should be able to integrate with your creative tools, e.g., Adobe Photoshop for images and Adobe FrameMaker for technical documents. Your DAM should also be able to share these documents seamlessly with your personalization technology.

Experience Optimization

Now that you have put together your Data, Content, and Personalization Strategy you can graduate to true one-to-one marketing. This is where you deliver on Experience Optimization utilizing your Data Strategy (Adobe Analytics) + Content Strategy (Adobe AEM Assets) + Personalization Strategy (Adobe Target).

To share how you deliver on Experience Optimization we will review how you utilize Adobe AEM Assets and Adobe Target to provide true one-to-one personalization. There is a third solution you would use as well for Experience Optimization (Adobe Analytics); however, that integration is an article unto itself – which very conveniently you can find here – Data-Driven Hyper-Personalization.

Adobe Experience Manager Assets

Adobe AEM Assets is an enterprise-grade DAM which allows you to store, deliver, and share among your team, images, video, and other content for any screen or device.

As a proxy for AEM Assets,  you will use the free Adobe Assets Core Service you have access to as an Adobe Target practitioner, which sufficiently supports our Experience Optimization use case.

We won’t delve into what is available in the paid AEM Assets version versus the free version (it is significant); however, here is additional information on the paid enterprise version of AEM Assets.

You will use the Adobe Assets Core Service to store the images you will later utilize in Adobe Target. As you can see, you are able to access Assets within the Experience Cloud menu.

When you click on Assets you can see the Folder and Files you have uploaded, and create a new Folder and upload new Files if you desire. You can also download, email, annotate, and add metadata to your images as well. Adding metadata is very helpful in the enterprise version which uses AI to automatically add metadata to images where there is none, and find images based on algorithmic image recognition. Find out more here.

Once you upload your images and add your metadata to your images, you will jump into the Target UI (which we will cover below) and add an image element to your page that you would like to test/optimize.

Adobe Target

You are going to use Adobe Target (Adobe’s personalization technology) to identify which new experience you create performs the best for your visitors. In other words, what can I show visitors to my website that will drive more leads/enable my company to make more money?

Your Challenge

To get to the ‘Request A Quotation’ submission form today, your visitor may have to visit several pages and serendipitously find your form. Not ideal. You realize this, and you decide to add the ‘Request A Quotation’ submission form to your homepage. But you are not sure where it would perform best. Therefore, you create three different homepage experiences and deploy a very specific type of A/B test to identify which homepage experience is performing best and let Adobe Target algorithmically drive traffic to that experience. To be clear you are not solely executing a standard A/B test to see what experience is performing best -– you’re executing a very sophisticated testing and optimization process to drive more visitors to the ‘Request A Quotation’ form.

The specific type of A/B test you will use in Adobe Target is the Auto-allocate A/B test. There are several different Adobe Target Activity types which can be found here. You chose the A/B Auto-allocate given the ease of test design and the powerful algorithm that drives optimization.

An Auto-Allocate A/B test identifies a winner among two or more experiences and automatically reallocates more traffic to the winner to increase conversions while the test continues to run and learn. It is costly to use a standard A/B test because you have to spend traffic to measure the performance of each experience and through analysis figure out the winning experience. Traffic distribution remains fixed even after you recognize that some experiences are outperforming others.

Your Solution

Below is where your ‘Request A Quotation’ form is located today. It isn’t easy to find at all. It is your main key performance metric and it is hidden on your website.

You are certain you can drive more RFQs if this element is located on your homepage so you ask, “where do we place the request on our homepage? And what does the request look like?”

Let’s walk through the steps to execute your Adobe Target Test.

Open Adobe Target and choose Activities > Create Activity. Here you are going to choose an A/B test. For an explanation of all of the various activity types, you can click here.

Within your A/B Test, you will be prompted to create your experiences. You decide that you’ll create three different experiences with three different calls to action elements to see which one works best and allow Adobe Target to algorithmically drive traffic to the highest-performing experience to optimize conversions.

Experience A

Experience B

Experience C

Now that you’ve created your experiences, it is time to append your Audience and choose your Traffic Allocation Method. If you recall when reviewing Data Strategy we discussed creating high-value audiences. For your test you want to understand the most efficacious experience among the three, so you’ll expose the different experiences to anyone who visits your site. For a very detailed article on the benefits of integrating Analytics audiences with Adobe Target, you can check out this article, Data-Driven Hyper-Personalization via A4T.

As we discussed earlier you will use a specific type of A/B test called Auto-allocate.

An Auto-allocate A/B test identifies a winner among two or more experiences and automatically reallocates more traffic to the winner to increase conversions while the test continues to run and learn. In our use case, the Auto-allocate A/B test will begin by initially dividing the traffic distribution evenly; however, as the algorithm gathers more data to learn from, it will begin incrementally increasing traffic to the experience that is performing the best. It is a combination of performance and statistical confidence that drives the algorithm. For more information on how the Multi-Armed Bandit algorithm functions in an A/B test, you can click here.

Now you are going to enter the success metric you’re looking to optimize. In this case, your conversion metric is the element that you created in your three experiences (‘Request A Quotation’). Once you click ‘Save & Close’, your Target activity is ready to run!

You can periodically check on the status of your test under Activities > Activity Name. Your activity will be either declared a winner or the test will still be in progress, and Target will let you know either disposition.

From the activity you can then click on ‘Reports.’ The reports section will provide you with a good deal of granular statistical insight into how your activity has performed to date (whether it is complete or not). Below is an overview of the type of conversion data you will receive:

Conversion Rate

The bolded number to the left is the median conversion rate and the number to the right is the number of conversions. There’s a bit more to this metric, but it’s a lengthy explanation. For more details click here.

Average Lift

The large number and arrow reflect the expected value of the lift. This number is the midpoint of the range of the lift bounds. The expected lift arrow displays in grey until the confidence passes 95%. After this threshold, the arrow displays as red or green based on the negative or positive lift.

Boxplot graph

The boxplot graph in the Target interface represents the expected value and 95% confidence interval of the success metric in question. The length of the boxplot bar represents how large the confidence interval is in an easy-to-understood way.

Confidence

The confidence of an experience or offer represents the probability that the lift of the associated experience/offer over the control experience/offer is “real”. Typically, 95% is the recommended level of confidence for the lift to be considered significant.

You should now understand how to run a very easy, yet very powerful Adobe Target A/B test. There are different activity types you can use for your tests that work through different use cases, however, for a beginners test in Adobe Target an A/B test using Auto-allocate is an excellent way to start, and this activity type will provide you with a great deal of confidence as to which compelling experience you should serve your customer. Adobe Target, when combined with Adobe Analytics, Adobe Experience Manager Assets, and the Adobe Creative Cloud, is the most powerful Experience Optimization Platform available today.

If you would like a demo of the Adobe Target, Adobe Analytics, and Adobe Assets solutions, either as point solutions or in the context of an integrated, multi-solution Experience Optimization platform, please reach out to our team. We’d love to help.