Blog

The case will be short, but interesting and clear.

You will see what A / B testing (split testing) is in contextual advertising with a simple, understandable example, and also what result this testing can give.

Case "A/B tests in contextual advertising"

We have a search advertising campaign in our account. As always, our main task is to attract a lot of conversions (sales) to the site and spend a minimum of money. There are a million ways, but in this case, we will test two different bidding strategies in Google Ads and see which one works best.

The original ad campaign uses a manual bidding strategy with a CPC optimizer. Create a duplicate of the original advertising campaign and set up another strategy in it - the target cost per action.

Don't be intimidated by the abstruse names of these strategies - just look at the table and the difference in numbers, and a description of the strategies can be found in the Google Ads Help.

Both campaigns work in parallel - one user is shown ads from the original ad, the other - an experimental one.

We are waiting for several weeks until a sufficient amount of statistics is accumulated for correct conclusions (remember the representativeness of the statistical sample).

Here's what we get:

Original advertising campaign versus Experimental advertising campaign:

• has a higher CTR (10.42% vs. 9.69%) - that's good!
• has less cost per click ($ 1.15 versus $ 1.62) - it's super!
• was shown more often in Search (% of impressions 73.22% versus 50.53%) - that's good too!

But with all this, the Original advertising campaign compared to the Experimental advertising campaign:

• spent 1066 USD (experimental - $ 835)
• brought the client 29 conversions (experimental - 35 conversions)
• conversion price - 37 USD (experimental - $ 24)

Therefore, we apply the tested strategy on an ongoing basis and are glad that now we will receive 21% more sales and 35% less money!

Conclusions:

1. the tested strategy worked significantly better than the initial one;

2. never rely on indirect indicators alone! We focus on the main indicators - macro conversions (leads / calls / requests / sales);

3. we continue A / B tests, everywhere and always, at different levels of the structure of the advertising account!

And the fourth conclusion, the most important: numbers should answer all questions, no intuition!

All sales! :)