Introduction
A dynamic audio advertising campaign can contain thousands of personalized audio ads.
Version an Ad is a workflow built to help you make dynamic audio ads, using a 'base' audio ad that can be automatically edited and produced into thousands of versions with differing dynamic elements – known as 'tags' – such as store call-outs, product offers (and much, much more).
Let's get started.
Step 1 – Set Up
Like any AudioStack workflow, you'll need access to it as a user, first.
If you don't have access, ask your admin or an AudioStack account manager to enable the Dynamic Creative solution for you.
It will appear on the left side of your screen. Click on "Start Workflow" to begin.
You will be asked to provide a base ad. You can upload or choose your base ad from files by clicking "Select".
💡 A base ad is a fully-produced ad in which the dynamic tags are missing. You'll need such an ad to be able to use Version an Ad.
Next, click on "Select a Project". This will open the Projects folder. A project will be created, and you can save the versioning project for picking it back up later or to share with a colleague.
At this point you'll also have to name your Session. Then, click next.
Step 2 – Version
Now, let's build your versions. You can build up to two dynamic tags anywhere in your base ad.
To position where the dynamic tag will go in your audio, click the pencil icon on the right of the player and reposition the purple dynamic tag box over the sound wave (your base ad).
You can resize its length by dragging/pushing along the edges. When you're happy, confirm by clicking the checkmark on the right.
💡 The first tag will appear as a blue box. The second tag will appear as an orange box.
To change the voice that will read your dynamic tag, you can click on the icon of the person and associated image. This will open the AudioStack voice library.
There are also more detailed settings for your tag. Click on the cog to open those:
Tag name: Lets you name your tag
Layout mode: 'Layered' means your tag will live "in" your base ad and 'Consecutive' will add the tag to the end of your existing ad
Length: Length of your tag (you can manually set the length of the tag in seconds, or use the drag-and-fit box as described above)
Starting time: Determines when your dynamic tag will start playing
Script alignment: Determines if your tag starts right at the beginning, in the middle of, or at the end of your blue or orange tag box.
You will also see this pop-up window when you add a second tag. Follow the steps above to position it.
Each tag will live in one of the tabs, which you can switch between at any time.
Step 3 – Creating the content of a single tag
This is a great place to start before you build thousands of versions.
Click on "Ad-ID" to name your file and use "Tag copy" to write or paste your content.
The audio wave icon on the right (directly below "GENERATE") will create your ad.
Once created, you can play it and fine-tune the wording or position of the tag.
Step 4 – Creating versions at scale
Once you're familiar with the way tag creation works, you can start to scale your production.
Common use cases for versioned ads include store location callouts, product offerings or legal tags, but there are many other things you can do.
💡 AudioStack also supports real-time data insertion (think stock market prices, sports results, or retail store inventory). This requires the use of our API.
You'll need all of the information before you start, and you can use the dynamic tag information in a spreadsheet format like you may already be used to.
However, we make a version available that works from the get-go. You can download the AudioStack versioning template (csv file) in Version an Ad. Simply click on the blue link saying "Template" to download the empty file.
Then, open the document on your computer and paste the right information into the correct columns:
Once you have filled in the information, save it to your computer.
💡 Make sure to save your template spreadsheet as a csv file, comma-separated so that Version an Ad can read it.
Next, return to the Version an Ad window in your browser. Click the "Import" button again and click "Choose file", where your file explorer will open. Pick your .csv file.
You should now see your versions in Version an Ad. They will follow the voice and timing parameters you established earlier:
Production menu controls
It's good practice to test your first tag before creating all of your versions. Simply click the voice icon (the little waveform) under "GENERATE" and take a listen.
Make changes as required and when you're happy with the output of the first version, click the "Generate" button in the upper menu. This will bulk produce all of the versions in your roster.
They will produce in batches so you can listen to your first version while the others are still rendering.
The production menu shows these buttons:
Generate - Generates all files at once
Import - Input your versions as text (or download the template).
Download All - Downloads all of the produced versions to your computer
Download Approved - Downloads only those files that have been manually approved
Create - Creates a single new version line
In the same menu, you will also be able to see how many files are in your project, how many have already been produced, and if there have been any files that couldn't be generated:
Once you generated all of your files, your screen should look something like this:
Use the player on the left-hand side to assess how your versions sound. You can click directly on the blue and orange tags to only listen to the dynamic parts.
All tags will be automatically amended to fit perfectly into your designated space – this is necessary as your dynamic tags will vary in length based on their content.
The three little dots on the right-hand side give you the option to duplicate, download a version as .wav, or delete a versioned asset:
Step 5 – Checking and approving your versions (QA)
You can use the QA boxes to individually check each version or share it with a co-worker asking them to have a look.
This process makes QA'ing multiple assets much more time-efficient. However, the most common issue you might run into relates to the speech in the tags being fixed: where the read comes out unnaturally because the tag length is too long or short for all the words that need to be read.
If you need to make a change because you're unhappy with the voice's delivery you'll need to change the words in the tag. This can be done by exchanging a period with an exclamation mark at the end of a sentence, adding a comma or rewriting a word.
This is quite rare, and as speech models are getting better and better, this problem will be even less of an issue.
🛠️ We're continuously working on improving the scalability of version creation, especially for those cases where you'll need several thousands of versions and human-QAing isn't feasible.
How to resume your work
If you need to take a break during the versioning process, you can easily stop your work in Version an Ad so that you can get back to it later. It will save automatically.
You can also make it easy for a co-worker to take over. Just go to "Projects" on the left-hand side menu, choose the name of the owner, open their project and right-click on the one you'll want to open:
