Visual quality control for advertisements.

Visual quality control for advertisements.

A core platform feature designed to be future proof & flexible enough to handle the processes of any team.

A core platform feature designed to be future proof & flexible enough to handle the processes of any team.

Company

Choreograph

Role

Product Designer

Date(s)

May 2024 - Present

The problem

The problem & approach

Choreograph create is a platform that allows users to create advertising campaigns from briefing to serving.

The problem at hand is that the platform currently lacks a quality control process, preventing teams from performing visual quality checks on their designs.

Our goal was to integrate a quality control feature that ensures all teams can consistently review and maintain the visual standards of their designs within the platform.



After some user interviews, my main focuses were the following:

  1. Give all users a clear overview of what state their advertisements were in the pipeline.

  2. Separate the process into two distinct steps

  3. Prevent users having to copy / paste hundreds of names and comments into external tooling.

  4. Let users bypass the process entirely.

This addresses two main issues brought up in the research:

  1. Prevent users having to copy / paste hundreds of names into external tooling.

  2. Advertisements in review will have their status updated in platform, so everyone involved in the project can know what stage each variant is at.

Impact.

Potential clients

Multiple major clients are waiting for in platform QC to sign up.

3

Existing agencies are currently beta testing.

Designs.

Providing a clear overview.

I needed to make sure the users have a good overview of the status for each advertising variant they are working with, as currently there is no way to tell if something is in review unless you have received an email telling you so.



When an advertisement is put into review, the status automatically changes from draft to review and depending on the recipient, the review status would indicate whether it is in an internal or client review stage.

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Columns with each reviewers status appears on the overview table once approval requests have been created.

Before avatars are introduced, a tooltip gives users more detailed information.

Easily filter ad variants based on their current status.

Easily create + assign an approval request

One, two step

During the discovery phase, it turned out that the majority of our agencies have their own unique processes for quality control, with a varying number of steps.

One of the key findings was that regardless of how many steps, or what the process is, there were two major steps that were always shared - the internal quality control process and the client review.

The biggest differentiator between these two distinct steps was that the client review was a simple final sign off based around a more professional transaction and the internal process was a little chaotic in nature, with varying steps per agency.



Diving the two steps has the advantage of preventing the clients from seeing any internal comments that may be deemed unprofessional and the internal team have full control of what the client sees and when.

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Reviews are split into two categories, internal review and client review.

Reviewers added determine the review type.

Manage access and permissions.

Two important edge cases were raised and prioritised; when a request is sent to the wrong person(s) & when you need to speed things up and want to add new reviewers to the process.



This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Easily manage reviewer access. This unsubscribes the user from further notifications.

Moving communication inside the app.

A lot of time is being spent by users communicating about the advertisment variant status in external tooling, including emails, chat applications e.g. Slack and project management tools, e.g. Jira. The tools tend to differ per agency, but the processes remain similar.



Often things get lost in translation and moving this functionality inside the application removes most risks of miscommunication and speeds up the entire process for all of our users.

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Fast one click approvals that automatically notify the users in app, in email and on the overview.

Comments were moved inside the application and scoped to each ad-variant to easily keep track of comments and changes.

Keeping the workflow flexible.

Agencies have different workflows, sometimes it can even differ per campaign. I updated the status lifecycle of avertisements by removing the approved status and changing the reviews to be advisory, rather than a compulsory step.



Although not recommended, sometimes desperate times require desperate measures. For when there is high pressure and little time, users can skip parts of, or the entire QC process by updating advertisement status' to ready.

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Despite having two rejected requests, a user can still move the variant along in the workflow.

Process.

Understanding the user needs.

I lead multiple workshops with users from the main agencies that use the tool. The workshops were held online as each team is based in a different country (UK, Netherlands, Denmark, Sweden & Columbia).



The output of the workshops was a Miro board with high level user journeys mapped for each agency, along with a list of the main issues faced when performing quality checks with their current methods.

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Key takeaways.

  1. Users need to know where each ad variant is within the workflow.

  2. Communication and handover moments need to be communicated within the app - whilst still allowing users to keep using their external project management tooling.

  3. We need to enable the option to fast track / skip the QC process for various use cases (e.g. time limitations).

  4. Users need separate rounds of quality control variants to their internal team(s) and their client.

Figuring out the workflow.


Several ideas for the workflows were explored ranging from Git inspired pull request pages to more integrated patterns used by our main competitors.



The later is what ended up sticking due to it involving less steps and it supporting the easy addition & removal of ad variants from active review requests.

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Competitor analysis.

Choreograph has several competitors in which I looked at their current workflows (Flash talking, Celtra & Hogarth) and documented / annotated on a Miro board.



Tooling outside of the typical ecosystem that involved the management and commenting of bulk assets was also looked at, one example being frame.io.

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Prototyping, testing & iterations.

I planned scenarios for user testing sessions, where I turned the screens into a functional prototype.

Four user test sessions were ran, where I asked users to complete certain tasks throughout the end to end process.

Several rounds of feedback sessions were also held with stakeholders and in total 6 iterations of the design were created, where each iteration resulted in a more simple workflow.



These user test tasks included:

• Asking the internal team to review specific variants
• Adding / removing new participants to the review
• Asking a client to review variants that have been successfully reviewed internally
• Approving variants
• Rejecting variants with comments
• Approving rejected variants which have had all of their issues resolved

This was initially scoped out of the MVP until checking with stakeholders, who raised the priority due to the potential sensitive nature of these types of mistake.

Solving problems since 2010

© 2024

Solving problems since 2010

© 2024

Navigation

Contact