A Quick Guide to A/B Testing

Digital advertising is much more efficient and relatively cheaper than many other forms of traditional ads. Determining which ads work and when to drop an ad is done by using A/B testing. Yes, you can test more than two options at once, but this the simplest form to get the point across for demonstration purposes.

With these testing methods, you can create two competing ads in various formats – based upon clicks and conversions you will be able to kill one ad and keep the ad that works best:

Testing Colors

Design the same ad in different colors – same images, same verbiage, but maybe the background color is different or the image colors are different. This works well with simple images. Load both and run them in the same manner – within a short amount of time you will be able to view the ad that gets the most clicks and converted the customer to your call-to-action.

Testing Content

Usually same images, same colors but the verbiage is different. Some ads just change the first few words to all caps and that is eye catching enough. Once again, this is also an effective way to quickly assess which content is preferred.

Testing Images

We are a visual society – using the same colors and verbiage but changing the image often results in a preferred ad response. You can quickly drop the underperforming ad and continue with the optimal ad image. Using the correct visuals can really boost conversion.

Ultimately your A/B testing should prove some analysis for future ads. This is the quickest and cheapest way to see what appeals without blowing your entire marketing budget.

Qualitative vs Quantitative Research

I teach “Principles of Marketing” at night at our local community college. I love teaching. analysis

The chapter that seems to cause the most problems every semester is the chapter that covers marketing research. The students always get stuck on what is qualitative vs. quantitative research.

So let’s make this simple.

Quantitative research implies that there is the ability to somehow “quantify” the results of a question.

This means you can use anything with a number as the response:

Quantitative research implies that there is the ability to somehow “quantify” the result of a question.

This means you can use anything with a number as the response:

    • Likert scale“On a scale of 1-5 with 1 being I don’t like it and 5 being I like it a lot – how do you rate the new flip top lid on our BBQ sauce”
    • Anything that can be measured (measurements, quantity, age, temperature, etc.) – “Please pour out the normal amount of our BBQ sauce you would use on your chicken breast after it has been cooked” (measure the sauce in ounces)
    • Yes/No questions (because they are finite and you can assign a number to it: Yes = 1, No = 0) “Would you buy the redesigned BBQ sauce with the squeezy bottle and flip top lid?

 

 

Qualitative research is “touchy feely,” more about “quality” or perceived quality. This type of research uses open ended questions that allow the respondents to answer openly without restrictions.

Questions such as:

  • “Are there other enhancements to our packaging of our BBQ sauce that you would like to see?”
  • “When you tasted the BBQ sauce sample A – what did you like/dislike about the taste of sample A?”
  • “Please taste the cooked BBQ chicken breast we cooked on a grill. Describe the texture and taste as a grilled BBQ sauce.”

If you have trouble remembering which is quantitative and which is qualitative, the quickest way to remember is that quantitative is about numbers and qualitative is about quality.

Keywords: #marketingresearch #research #qualitative #quantitative