1. Text analysis
So far, we've reviewed marketing levers and challenges with marketing data overall. Now, we'll review different analysis approaches, starting with text analysis.
2. Text analytics basics
As Marketing Analysts, we need to analyze text as part of channel performance.
Text analytics is not just analysis that involves text, it is taking unstructured text and translating it into quantitative insights.
If we think about various channels, text is included in the majority of marketing ads (either spoken or written).
Paid search ads are the most common example - most ads are entirely text-based.
Social media also contains text in ads, as well as in comments.
Even radio and TV have text in ads with voiceovers or messages.
Text analytics can be limited to a single channel, span multiple channels in a campaign, or include the entire marketing program.
3. Paid search keywords
Let's start with paid search since that is the most common form of text analytics.
Paid search ads rely on the terms users type into search engines, called "keywords."
Marketers pay to have ads appear when someone searches for a keyword.
Keywords fall into two main buckets: branded, meaning, with a brand name, or non-branded meaning, no brand name is mentioned. For example, if our brand was Coca-Cola, "new Coke flavors" is a branded keyword, but "12-pack of soda" is a non-branded keyword.
Because many brands compete for non-branded keywords, this makes them more expensive generally.
Branded keywords are less expensive, but their volume is tied to how well-known the brand is. Marketing Analysts need to keep these nuances in mind as they analyze paid search data.
4. Paid search analysis
Another factor to consider when analyzing paid search keywords is that marketers set targets to keep costs from becoming too high.
It is not enough for analysts to analyze cost by branded and non-branded keywords; they must compare to cost per click targets.
Cost per click, or CPC, is basically taking the average of all keyword costs divided by all keyword clicks.
Once we learn the targets for paid search, we can evaluate dimensions to see when we paid more or less than our target.
In this chart, we can see that LATAM (short for Latin America) keywords are usually below target, while keywords in EMEA regions (or Europe, Middle East, and Africa) are above CPC target.
Analysts should monitor CPC trends to make sure that costs are not increasing over time and hurting overall channel performance.
5. Cross-channel text analysis
Marketers would like to see text analysis that is cross-channel, which has a larger impact since it can cover the entire marketing program.
Analysts can manually classify data if it is a relatively small dataset.
However, for larger data, we need to consider natural language processing or NLP.
NLP is using computer programs to make sense of language as it is spoken and written.
This means NLP can work with spoken ads (like in radio) and find similar categories with paid search ads.
Eventually, analysts could compare text by theme, agnostic of spelling mistakes and language differences, across all channels.
6. Let's practice!
Time to try out paid search analysis concepts!