Contemplating if your audience will enjoy the film or image, or whether a questionable or straightforward title is best? We’ve all been there when crafting Facebook ads. To guarantee that your adverts receive maximum attention and success, A/B Testing provides an invaluable solution.
What is A/B Test?
By definition, A/B Test is an experiment that evaluates and contrasts different advertisement elements to decide which performs better. For example, you have crafted a version of an ad but think another headline may be more successful; by setting up two versions of the ad through AB Testing, you can distinguish between them and find out which works best. As demonstrated in the following case study where only one banner emoji was included in a title - this increased click-through rate drastically by 50%.
The importance of AB Test
Facebook's AB Test is an exemplary tool for quickly and precisely assessing the efficacy of advertisements. Instead of relying on guesswork, it utilizes a direct approach to compare different ads, enhance CTR (click-through rate), and minimize conversion costs with unerring accuracy.
How to set up AB Tests on Facebook/Meta Ads?
When carrying out an AB Test in Facebook Ads, you must adhere to a distinct three-step sequence: decide which variable to test; set the relevant elements; and assess the results.
Facebook ads are divided into three tiers - Marketing Campaign, Ad Set and Ad - with different testing variables at each level.
At the Marketing Campaign stage, it is essential that you pick your target audience first since this will define what template formats and placement optimization/bidding alternatives your ad portfolio can have access to.
It is not recommended that one performs an AB Test across various campaigns as there are more considerable options for testing within individual ones instead.
Ad Set Variables:
- Advertising Optimization Method
- Ad slot
- bid strategy
- audience choice
If you want to test different ad combinations, make sure each ad combination contains similar ads, otherwise you won't get relevant test results.
Adespresso spent 3 million advertising expenditures, and summed up several key variables that affect the advertising effect the most:
- nationality
- precise interest
- Mobile OS
- age range
- gender
- picture
- job title
- relationship status
- list collection page
Audience targeting factors and ad portfolio elements can be tested for efficiency by creating multiple portfolios with individual versions of the element being examined. Utilizing A/B testing, you'll easily discover which version works best.
Ad variables:
- Ad type
- Media (photos vs videos)
- Content
- Title
- Link description
- Call to Action
Facebook/Meta Ads AB Testing Best Practices
1）Test only one variable
By testing just a single variant, marketers can quickly and easily analyze the results.
2）Use the same ad structure
Let's say you have 2 headlines, images and pieces of content to test. Placing them in the same ad group will cause Facebook to optimize these ads automatically, leading to a competition between tests with different objectives. As such, results may be difficult or nearly impossible to distinguish; thus their relevance is quite low too! A better strategy would entail creating 3 separate groups for each one of these variables and testing them individually - this way your outcomes are clear-cut and accurate!
3）Make sure the test results are meaningful
Obtaining the appropriate quantity of data is essential for obtaining results that are statistically significant. Fortunately, there are numerous statistical tools available to assess the reliability of statistics; abtestguide's tool is one such resource - it suggests collecting more than 100-500 conversions or clicks per test in order to guarantee accurate outcomes without needing any additional statistical analysis.
4）Make sure you have enough ad budget to make the results meaningful
As Adepresso's statistics reveal, running an AB test with 4 variables requires a budget of around US$100 per variable, which makes a total of US$400. It is possible that the testing time and cost could be reduced if certain tested variables are performing exceptionally well; however, spending too much to run a test may not always lead to favorable outcomes—it is best practice to select several key elements within your budget limit and allow sufficient funds for thorough experimentation.
Analyze the results
Ad Manager will alert you of the top variants in your A/B Test. For instance, when it comes to Lead Generation campaigns, the most successful version is determined by Cost Per Lead (CPL) — Facebook keeps an eye on which variation has the least expensive CPL cost.
Additionally, you can measure CPC and CPM numbers; however, they will not be able to tell you exactly how much it costs to get a conversion. To ascertain the maximum reach of your branding campaign after an ad has been running for some time, consider metrics as well.
If you're interested in improving your Facebook/Meta ads skills, subscribe to our blog to stay ahead of the competition!
觀眾到底是喜歡影片還是圖像？選擇的標題是疑問句還是直述句？這些問題是我們曾經創建Facebook廣告時都會遇到的問題。如果想要確保Facebook廣告可以發揮最大成效，A/B Test就是關鍵能功,令你能抓住觀眾的眼球。
什麼是A/B Test?
簡單來說，A/B Test是驗證廣告成效的假設。系統通過測試不同的廣告元素，找出哪個廣告表現最為出色。
例如你設計了一個廣告的版本之後，覺得另外一個標題可能表現更好，為了驗證你的假設，你需要進行AB Test來設定2款廣告，來觀察哪一個廣告表現更好。
下面2個版本的廣告只不過標題多了一個旗幟的符號，點擊率就增加了50%。
AB Test重要性
Facebook廣告AB Test成效顯著因為不使用主觀邏輯來推斷廣告成效的手法，而是採用直接客觀的方式來比較不同的廣告成效，提升CTR（點擊率），從而降低轉化成本，整體來說能節省下來的廣告費的預算是非常可觀的。
Facebook廣告AB Test是怎樣設定？
Facebook廣告裡面的AB Test設定需要遵循以下三個順序：選擇要測試哪一項變數；設置相對應的項目；分析測試的結果。
Facebook廣告架構分為三層：行銷活動Campaign，廣告組合Ad Set和廣告Ad，不同層級有不同的測試變數。
行銷活動Campaign：首先需要選擇目標，你的廣告目標決定了你的廣告組合具有哪些廣告模板和投放優化和出價選項。
通常來說，AB Test不會在不同的行銷活動下進行，因為在廣告組合和廣告中還有更多實質的可以用來測試的選項。
廣告組合Ad Set：可以測試的變數包括：
- 廣告 優化方法
- 廣告版位
- 出價策略
- 觀眾選擇
如果你想測試不同的廣告組合，請確保每個廣告組合都包含相似的廣告，否則無法獲得相關的測試結果。
Adespresso花了300萬的廣告支出，總結出來幾個重點影響廣告效果最大的變數：
- 國家
- 精確的興趣
- Mobile OS
- 年齡範圍
- 性別
- 圖片
- 標題
- 關係狀態
- 名單收集頁
這些受眾群體的定位因素和廣告組合元素可以通過創建多個廣告組合進行 A/B Test，而每個廣告組合則包含被測試元素的不同版本。
廣告Ad
可以測試的變數包括：
- 廣告類型
- 圖片 or 影片
- 文字
- 標題
- 連結描述
- Call to Action
Facebook廣告AB Test最佳實踐
1）系統只測試一個變數
系統只測試一個變試，可以令行銷人員更容易追蹤和評估效果，使用的變數越少，你可以在更短的時間內確認結果。但若是一口氣火力全開全都丟下去測試，就會發生 …
2）使用正確的廣告架構
不要將所有的變數放入同一個廣告組合，例如，你有2個標題、2張圖片、2個內文要進行測試，應該將這三個不同的內容放在三個廣告組合分別測試。如果將這些標題、圖片和內文放在同一個群組一起測試，Facebook會自動優化這些廣告，這些不同測試目標的廣告將會一起競爭，最後的結果會很難分辨各個變數效果的好壞，得到的結果相關性是很低的。
3）確保測試結果是有效的
你必須得到一定的數據量，才能夠得到統計上顯著的測試結果。有一些統計工具可以檢測數據的信度，例如:abtestguide的這個工具假設不使用統計工具，則每一次測試，最好能夠蒐集超過100~500次的轉換或是點擊。
4）足夠的AB測試預算
根據Adepresso數據統計，如果你的每一個轉換是US$2.5，而你需要測試的變數有4個的話，預算是$2.5 x 4 x 300 = $3,000。 但如果你的測試變數中，有的變數比其他的明顯表現更好，會縮短測試的時間，也會降低成本。 但一般而言，除非你有無限的預算，否則不建議過度進行AB測試。建議挑選幾個變數，給一個合理的預算進行測試。
分析測試結果
Facebook 會在Ad Manager中通知你A/B Test得到更好結果的廣告版本。不同的Campaign目標就會有不同的衡量準則，例如“開發潛在顧客”（Lead Generation）的Campaign，會以每位潛在顧客取得成本（Cost Per Lead/CPL）來衡量勝出的版本（最低CPL的版本勝出）。
而你當然也可以使用 CPC、CPM 來衡量，只是它們可以無法明確說明 "需要多少成本才能得到一個轉換數" 。另外在廣告運作一段時間後，也可觀看像「進行品牌宣傳活動，能獲得最大的曝光數」這類指標。
Runner, singer, digital marketer, and entrepreneur. I'm always excited for the future!
.jpg)