[PDF] Adobe Summit 2018 TechLab 12 LabGuide

View - Download Adobe Summit 2018 TechLab 12 LabGuide

Previous PDF

Next PDF

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

TL12 - Target QA & Debugging

Introduction

Instructors

Thomas Heinbach - Principal Solution Consultant, Adobe Systems Abhinav Sharma - Senior Analyst Value Audits, Adobe Systems Introduction and Lab Objectives Doing QA for Target is super easy! That's what those preview URLs are made for, right?

Well, yes - in a way.

Having those preview URLs makes life for CRO-teams easier for testing the different versions in a test. But more and more optimisation activities also include different versions for different visitor segments based on dynamic data and/or personal visitor profile data, for which you'll need many, many preview URL's. On top of that: mobile responsiveness, adaptive websites and the increasing diversity in device types, operating systems and browsers have added more complexity to the QA process. Furthermore: the increasing adaptation of new technologies like server-side targeting, Target in emails and Target in mobile app s requires more "non-standard" QA processes. In this session, we'll share QA-best practices that will help you be a more successful optimisation organisation. Lab Developer Resources This lab will be presented as an interactive session during Adobe Summit in London on May 3rd 2017.
A demo website and this guide is set up at http://summit2018.adobedemo.de/ .

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Select your Level

Level 1 - I am a non-technical business user (marketer, optimization manager, consultant, mobile product manager etc) and I am not familiar with Adobe Target or the concept of A/B Testing. Level 2 - I am a non-technical business user familiar with Adobe Target or the concept of A/B testing and I have used it on the web. Level 3 - I am a developer, architect or implementation consultant.

Guidance

!If you are in Level 1, you may be a bit overwhelmed by the level of detail we are providing in this lab . You are welcome to sit back and enjoy the show though. However, you may consider passing on your place at the keyboard to somebody who is in Level 2 or 3. !If you are in Level 2, many of the things we discuss in this lab will look technical at first but in fact are easy to do. Enjoy the lab! !If you are in Level 3, you may know a lot of the things we teach in this lab already. Feel free to share your insight with all of us when you think it's appropriate.

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

What could possibly go wrong?

Concurrent Activities leading to conflicts/collisions An activity collision occurs when multiple activities are set up to deliver content to the same page. If an activity collision occurs, you may not see the expected content on your page. If your activity contains potential collisions, the Collisions tab is available the on the activity overview page. All activities on the same URL are listed, regardless of any audience targeting

in each activity. Open this tab for a list of activities that might be colliding. Click an activity in

the list to view the overview page for that activity. If the collision alters the expected experience, edit the activity.

The Collisions list helps you:

Identify whether a test is already running on a page before you set up a new activity Troubleshoot an activity if the expected content does not appear The Collisions list shows every Target Standard scenario where the mbox is used and that uses the same URL. For each potential collision, the list shows the Activity URL, the mbox name where the collision might occur, and any activities that match bot of those criteria. If there are multiple mboxes, they are each listed. The list shows the status and priority of each potential collision, along with other information. You can use the status and priority to help you determine the likelihood of a collision occurring. For example, if there is a potential collision between two activities and one is

inactive, there will be no actual collision unless the inactive activity is activated. If the potential

collision is between two live activities with the same priority and the same audience, a collision will occur. You can change the priority or status to prevent the collision.

Note: if

an unresolved collision happens (i.e. two activities with the same priority are using the same mbox on the same URL) Target will decide which activity to execute by looking at the last activation date. The last activity that was activated wins. If the audiences are different, there is still a potential collision bec ause it's possible a particular visitor could belong to multiple audiences.

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Segment Mismatch

Problem 1: You don't qualify for the audience targeting conditions. Validate: Review targeting conditions on the overview page and / or segment information as we show you later in this lab. Options: Qualify (set your environment in a way that allows you to enter the activity) and try again. Note: You can use Preview Links to bypass targeting conditions. Problem 2: The page doesn't qualify for the page targeting conditions. Validate: On the overview page, determine if the page falls outside of the targeting conditions. Options: Go to the Visual Experience Composer, click URL-->Advanced-->current page.

Unexpected User Profile Values

Problem: Very often Segments are defined via user profile values. If the user profile does not fit the expected segment, you cannot see the activity experience you would have expected. Validate: check the user profile data and segments as we will show you later in this lab. Options: Qualify (set your environment in a way that allows you to set the profile data needed to enter the activity) and try again.

Mbox Timeout ! "Disabled"

When a Target servercall (often nicknamed "mbox call") does not receive a response until it times out, Target will go into "disable" mode and no longer try any servercalls for 30 minutes. Mbox.js sends a cookie called "em-disabled" to the visitor if target.js fails to load during delivery. This cookie prevents offers created using the Visual Experience Composer from rendering on the site. Visitors with this cookie neither see the test content nor get counted in those activity reports. All other offer content (from campaigns in Target Classic for example) continues to load. The cookie has a lifetime of 30 min from the time of load failure Problem: Your Target servercall timed out and Target won't make servercalls for 30 minutes. Validate: check the mbox cookie. If it has the value "disable" somewhere, Target is in disable mode. Check if a cookie "em-disabled" exists. Options: !Check your network settings for reasons why the response may be slow. Normal response times are in the range of 20-120 milliseconds. !Delete the mbox and/or em-disabled cookies and try again.

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Standard QA Procedures

Target Default Publishing Workflow

Activity QA

Preview URLs & Best Practices

Use a different browser for working in Target and for QA

OR: Use the incognito window

Set up your QA browser in a way that it deletes cache & cookies whenever it is restarted Don't delete cache & cookies on your work browser, or you'll have to log in again.

Use a supported browser (! release notes)

QA Mode is "sticky" - if you use the same browser, you may need to release your browser from QA mode: bookmark.html

Role Description

Observer Can view activities, but cannot create or edit them. Editor Can create and edit activities before they are live, but cannot approve the launch of an activity. Approver Can create, edit, and activate or stop activities.

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

QA best practices

Test the look and feel for mobile web in VEC Viewports

Mobile viewports help you

preview how your activities appear on screens of various sizes.

The mobile viewport preview

feature is designed for responsive sites. Use mobile viewports if your site is responsive and the same elements in your desktop page are used on your mobile page in a different configuration.

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Standard QA Best Practices

Once you are done with the activity definition and enter the QA procedure proper, you will typically test the experiences you defined in your activity based on:

Browsers

OS screen sizes device types

Browser Emulators & Co

Firefox' Responsive Design View

Firefox has a responsive design view that encourages you to stop thinking in terms of specific devices and instead explore how your design changes at common screen sizes or your own size by dragging the edges.

Edge's F12 Emulation

To emulate Windows Phones, use Microsoft Edge's built-in emulation. Since Edge does not ship with legacy compatibility, use IE 11's Emulation to simulate how your page would look in older versions of Internet Explorer.

Commercial Emulators

BrowserStack (commercial) is the easiest to use for manual testing. You select an operating system, select your browser version and device type, select a URL to browse, and it spins up a hosted virtual machine that you can interact with. You can also fire up multiple emulators in the same screen, letting you test how your app looks and feels across multiple devices at the same time. SauceLabs (commercial) allows you to run unit tests inside of an emulator, which can be really useful for scripting a flow through your site and watch the video recording of this afterwards on various devices. You can also do manual testing with your site. Device Anywhere (commercial) doesn't use emulators but real devices which you can control remotely. This is very useful in the event where you need to reproduce a problem on a specific device and can't see the bug on any of the options in the previous guides.

Source: Google Online Help

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Enterprise Permissions

Properties & Workspaces

Enterprise Permissions are available in Adobe Target Premium only. So, what are properties and workspaces?

Property

Properties are similar in nature to those within Dynamic Tag Management (Activation) in that they use a unique snippet of code to differentiate them. A web property is a library of rules and one embed code. A web property can be any grouping of one or more domains and subdomains. Properties are enabled by adding a specific name/value pair as a parameter with any call (mbox, api, etc.) to Target. Properties belong to specific channels (Web, Mobile, Email, or API/Other).

Workspace (Product Profile)

A workspace lets an organization assign a specific set of users to a specific set of properties. In many ways, a workspace is similar to a report suite in Adobe Analytics. Note: Workspaces are known as Product Profiles in the Adobe Admin Console for Enterprise. If you are part of a multi-national organization, you might have a workspace for your European web pages, properties, or sites and another workspace for your American web pages, properties, or sites. If you are part of a multi -brand organization, you might have a separate workspace for each of your brands. Users can be part of multiple workspaces and can even have different roles within each workspace. Users can have different views of Adobe Target by moving between workspaces, similar to how Analytics users have different views of Analytics by moving between Report Suites. Workspaces can include complete different audiences, code offers, and activities.

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

All audiences and activities created before the new Enterprise Permissions model migration will be grouped together in the "Default Workspace," discussed below. All activities created via Adobe Experience Manager (AEM), Adobe Mobile Services, and Adobe Target Classic will be part of the "Default Workspace."

Default Workspace

All existing workspaces (product profiles) within Admin Console are merged into a single workspace called "Default Workspace" during your organization's migration to the new

Enterprise Permissions model.

Important: Do not delete the Default Workspace.

User Groups

You can create user groups, such as Developers, Analysts, Marketers, Executives, etc., and then assign privileges across multiple Adobe products and workspaces. Assigning a new team member all the appropriate privileges across different Adobe products can be as easy as adding them to a specific user group.

Roles and Permissions

Roles and permissions determine the access levels that users have to create and manage activities in your Target implementation. In Target, roles include the following: Observer: Can view activities, but cannot create or edit them. Editor: Can create and edit activities before they are live, but cannot approve the launch of an activity. Approver: Can create, edit, and activate or stop activities.

Channel

Channel refers to the content type of where your Target activities are delivered: webpages, mobile apps, email messages, and so forth. When you create a new activity, it is created in the currently selected workspace. You'll see channel selection options in the first dialog box that lets you choose the desired channel for the activity: Web, Mobile App, Email, or Other/API.

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Target Host Groups

With Target Hosts and Environments, you can organize your sites and pre-production environments for easy management and separated reporting. Similar functionality existed in Target Classic. Host groups in Target Classic are now called "environments" in Target Standard/Premium. The primary goal of host management is to ensure that no inactive content accidentally appears on websites. Host management also lets you separate report data by environment. A host is any web server (or web domain) from where you serve content during any stage of your project. Any host serving an mbox is recognized. Hosts are bundled into environments for ease of management. For example, you might have dozens of hosts grouped in two or three environments. The preset environments include Production, Staging, and Development. You can add new environments and rename your environments, if desired. One environment, the default environment, is pre-named Production. This default environment cannot be deleted, even if you rename it. Target assumes that this is where you will serve final, approved activities and tests. When an mbox request is received from new websites or domains, these new domains always appear in the Production environment. The Production environment cannot have its settings changed, so unknown or new sites are guaranteed to see only content that is active and ready. Host management also lets you easily ensure the quality of new activities and content in your test, staging, and development environments before you activate the activities. Target does not limit a host that can send and receive mboxes, so when new servers or domains come up, they automatically work (unless you've set up a whitelist or blacklist). This also enables ad testing on different domains you don't know or can't anticipate. Please note: Environments can be set up to execute inactive Activities, too!

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Debugging Target

Debugging Considerations

Cookies & mbox cookie settings

Watch out for 3

rd party cookie blocking browsers "XDomain" setting in mbox.js and at.js: 1 st , 3 rd or both ?

Mbox timeout

Default is 15 seconds (mbox.js), but should be somewhere between 2 and 5 seconds A servercall that times out will disable further servercalls for 30 minutes.

Priority setting in Activities

A way to avoid conflicting activities

"fine" vs. default setting "Manual" Debugging Techniques mboxDebug

It's d

epracated. Don't use it. If it still works, update your Target JS library. mboxTrace mboxTrace enables you to receive trace information attached to mbox replies. Trace information reflects the outcome of an mbox call (for example, a conversion or an impression) and any additional data that may help in determining why this particular outcome happened, such as a set of available branches among which the selection was made in a campaign. Use this information to debug content delivery.

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Example mboxTrace Call:

9619-xxxxxxxxxxx

The authorization key can be generated via Setup ! Implementation (scroll all the way to the bottom) The output displays very detailed information about your content. mboxTrace shows details about your campaign or activity and profile It also provides a snapshot of the profile before execution, and a snapshot of what changed after execution. It also shows which campaigns or activities were evaluated for each location. ./"01)2($%3+,*"-'%34,(".$% ?mboxTrace=console

Prints into console log as objects

For at.js, instead of popping a new browser window or outputting to the console as in mbox.js, you will need to inspect the Network request and look under Preview (Chrome) or Response (Firefox). ?mboxTrace=json Prints into console log as a literal JSON string ?mboxTrace=window Prints into a popup window as a JSON string ?mboxTrace=disable Turns off tracing session mode

Adobe Summit London 3

rd /4 th

May 2018

TechLab 12: Target QA & Debugging

Some of the information includes matched and unmatched segment and target IDs: SegmentId: The IDs of segments, either from the reusable segments library or anonymous ones created for the particular campaign. TargetId: The IDs of targets, either from the target expression library or anonymous targets for any segments from campaign.

Unmatch

ed: The request did not qualify in this call for those segments or targets. Matched: The request qualified for the specified segments or targets.

Using mboxTrace on Recommendations pages:

Adding mboxTrace as a query parameter on pages with recommendations replaces the Recommendations design on the page with an mboxTrace details window, which displays in- depth information about your recommendations, including: Recommendations returned vs. recommendations requested The key used, and if it is generating recommendations Criteria-generated recommendations vs. backup recommendations

Criteria configuration

Exclusions and inclusions applied

Collection rules

You do not need to include =console, =json, or =window in the query parameter. When you are done with the mboxTrace details, add =disable and press Enter to return to the normal display mode. The normal functioning and appearance of your site is not affected by mboxTrace. Visitors will see your regular Recommendations design.quotesdbs_dbs17.pdfusesText_23

[PDF] adobe experience platform debugger firefox

[PDF] adobe experience platform etl

[PDF] adobe experience platform rest api

[PDF] adobe experience platform security

[PDF] adobe experience platform source connector

[PDF] adobe experience platform tutorial

[PDF] adobe experience platform vs adobe audience manager

[PDF] adobe experience platform white paper

[PDF] adobe experience platform xdm

[PDF] adobe extendscript toolkit 2020

[PDF] adobe extendscript toolkit cc 2017

[PDF] adobe extendscript toolkit cc 2019 download

[PDF] adobe extendscript toolkit cs5

[PDF] adobe extendscript toolkit cs6

[PDF] adobe extendscript toolkit cs6 download