Tanya Tsukrova Emma App Challenge
Exercise 1:
This table refers to the last 7 days campaign performance on Facebook / Instagram. Based on this data, please answer the following questions:
- How much did you spend on campaign 2? £19,221.98
- Which campaign has the best click-through rate? Campaign 3 - CA
- Provided you can split the budget however you choose countries and you have no volume constraints, which campaign would you prioritise to get the lowest cost per registration? Campaign 1 - UK
- If all purchases had the same value, which campaign would have the best ROI? Campaign 2 - US
Exercise 2:
This table refers to historical data for a specific campaign where ad sets are split by age & gender. Retention is calculated as the percentage of users who came back to the app after registration (Day 0). Based on this information:
- Which ad set has the cheapest cost per purchase? Ad Set 6 - Female 50-65
- Assuming cost per registration and retention are fixed, and you could put all the budget into one ad set to get as many active users as possible in 7 days, which ad set would you choose? Ad Set 2 - Female 20-35
- Which age group has the highest cost per retained user (30 days)? Ad Set 5 - Male 50-65
- How would you reallocate budget to optimise for registration?
There are 3 constraints:
- Total daily budget must stay the same
- Max budget of £2,000 per ad set
- You can only pause 3 ad sets
| Daily Budget | New Daily Budget | |
|---|---|---|
| Ad Set 1 - Male 20-35 | £500.00 | £1,300.00 |
| Ad Set 2 - Female 20-35 | £500.00 | £2,000.00 |
| Ad Set 3 - Male 35-50 | £2,000.00 | £1,300.00 |
| Ad Set 4 - Female 35-50 | £1,000.00 | £0.00 |
| Ad Set 5 - Male 50-65 | £250.00 | £0.00 |
| Ad Set 6 - Female 50-65 | £350.00 | £0.00 |
As ad set 3 had higher spend than ad set 1, there is a chance that with more spend ad set 1 will optimise performance further and have lower CPR in the long run, hence splitting the budget between the two evenly to see how CPR will be affected.
Question 3:
How do you come up with new creative ideas to test? Which creative types & themes work best on social platforms in your opinion?
I start by researching what tests have been done in the past by the company. Depending on how old those tests are, I decide whether it's worth starting tests from scratch to build the knowledge base or picking up where my colleagues left off.
In my experience, video creatives work best for Meta and TikTok for non-catalogue products. Dynamic creative ads perform quite well for BAU ads and can also provide some insight into audience preferences on creatives.
To explore themes, I work with the data team to research the main motivations for people to join the app and create marketing personas based on that. Themes are mostly defined by personas but can also be expanded with other creative ideas, such as seasonal or recent events and testimonials. When I am off work, I always keep an eye on the ads of other companies and use them as a ground for new inspirations.
Once I know which themes resonate with the audience most, I can start perfecting the creatives by running an array of tests, such as video vs. image, square assets vs. vertical assets, long video vs. short video, dominant color: purple vs. white, product featured vs. aspirational image, UGC vs. BAU, CTA A/B test, price vs. no price and many more.
While I do think certain creatives and themes work better than others, I believe it's always good to challenge that knowledge with A/B tests to ensure I deliver the most efficient creatives possible.
Question 4:
Which strategy do you employ when testing new creatives on live campaigns / ad sets?
I either duplicate an existing campaign, ad set, or ad and change a variable, or compare two existing campaigns or ad sets. I use the same budget for both versions in a test to make sure the comparison is fair. For tests, I normally go from the most impactful elements to the least impactful; for example, visuals will come first (video/image), then CTA and headlines, and then the main text message above the creative. People tend to skim through ads, so it makes sense to start with what makes the most impact.
Question 5:
Which audiences would you test with Emma and how would you allocate budget for testing new audiences?
The audiences that immediately come to mind are people who are looking into self-improvement and consume self-help content. Getting finances in order has a good overlap with the audience's age (late 20s to early 30s) and location-wise (UK-US). For this audience, affiliate or influencer marketing can be quite strong and would allow more acceptable pricing compared to Meta or TikTok campaigns. As for Meta and TikTok, I would be curious to see how lookalike audiences compare to automated broad targeting. Location-based targeting in the US can be quite beneficial too.
I prefer to allocate 15-30% of the budget to testing; however, the real number will depend on what the business prioritises at the moment.
Question 6:
How do you keep updated with the latest trends and industry news regarding Facebook/Instagram ads and campaign optimization?
- Meta newsletters and webinars
- ByteDance newsletters
- App Promotion Summit
- Catching up with colleagues in the field
- LinkedIn and Reddit
- Being chronically online
Question 7:
In your experience, what is the best FB ad account structure (how do you split campaigns, ad sets etc) and why?
Campaign structure is defined by two main factors: business goals and Facebook limitations. In the past few years, the push from Facebook has been to consolidate as many campaigns as possible to allow machine learning to find the best audience possible. However, in practice, a combination of warm and cold targeting may work best. This is something to be tested on the account. I would recommend having no more than three active campaigns at once:
iOS Campaigns
| Campaign | Status | Ad Set | Objective |
|---|---|---|---|
| Prospecting | Always on | Prospecting | Registration |
| Retargeting | Always on | Retargeting | Purchase |
| A/B test | On demand | - | - |
| Prospecting Promotion | On demand | Prospecting | Registration |
| Retargeting Promotion | On demand | Retargeting | Purchase |
Android Campaigns
| Campaign | Status | Ad Set | Objective |
|---|---|---|---|
| Prospecting | Always on | Prospecting | Registration |
| Retargeting | Always on | Retargeting | Purchase |
| A/B test | On demand | - | - |
| Prospecting Promotion | On demand | Prospecting | Registration |
| Retargeting Promotion | On demand | Retargeting | Purchase |
Ads would be determined by the type of campaign and budget. The number of ads should allow 50 conversions per ad set per week to exit the learning phase while also taking into account the available resources of the design/copy teams.