Ofer Zeevy
Win, lose or draw: the three choices in customer feedback as analyzed by our AI
Updated: Jul 7, 2022
If there's a word for it, we can track it. But then we will also categorize it, report on how many times it was mentioned in a specific time period, analyze it in and of itself but also with regards to the bigger picture. Finally, we will come to conclusions about the necessary steps for our clients to attend to and act upon. And it all starts with the words and mentions of the company's customers all over the open web.

Real-time and constant brand sentiment measurement, one out of many features in our Product Enablement platform, calculates and updates real-time positive vs. negative mentions. Words or phrases that do not fit any of those two categories fall under "neutral". Brand sentiment is then determined by the percentage of the positive feedback vs. the negative comments out of the total real-time conversations.
While the AI platform and machine learning are not perfect, our algorithm is constantly being taught and improved in order to present the most accurate sentiment analysis solution for brands. Sometimes sentiment can be subjective, especially with the most neutral types of conversations and Affogata allows people from our client companies to manually categorize each mention or feedback on the platform and change the sentiment according to their KPIs.
Clients often ask our team to explain how our AI platform decides on the words' categorization it tracks in real-time. While there are obvious positive or negative words that fall in their respective categories, our AI also must decide on words for the often-times puzzling third category of "neutral". The following examples will better explain our categorization process.
Positive
Positive comments usually include words and discussions which indicate that the customer appreciated a product or a service. Common adjectives used include words such as appreciate, fantastic, fun, best, awesome, amazing, love, smile, gorgeous and fabulous.
Descriptions with a positive tone include mentions about ease of use or thanking for a good service. A comment of how the product or service is user-friendly counts as well in this category. Referrals, that is, indications about how the user recommends the product to another are also considered positive. And finally, a love or a smiley emoji is also tracked here.
Neutral
The second category may baffle a little but it basically includes all of the mentions which are not clearly positive or negative. Comments, where people are not taking a stand or voicing an opinion, will fall under "neutral". Then there are those who are asking a customer service question while not favoring or criticizing the product.
Suggestions for improvements also count here, since they naturally present a mixed bag: the customer is apparently not happy with a product feature or part of the service (a negative sign) but he is also proactively devoting his time and effort to contribute to the possible improvement of it (a positive sign). So a mixed bag is categorized and included in the "neutral" zone.
Finally, mixed bags may also be reported when customers rate a product high (a 5-out-of-5) but may add a negative comment about a specific problem with it. Since their written critique balances their high score, their feedback falls under the category of "undecided" which is "neutral".
Negative
The third and final word category includes "all things bad". In the sub-category of "bugs", for example, words such as crash, glitches, and errors indicate negative feedback. Adjectives such as bad, hate, terrible, awful, and horrible fit neatly into this category as well. Then there is the category of the complaint with such mentions as "I can't close/access/open the account" or "I can't find the stuff" or "I tried the link but it got me nowhere".
Customers react negatively when they are also not happy with an app's tutorials, commenting on the subjects they still can't figure out. Physical or digital products may receive conversations such as "Your product is not working" or "It has poor quality". Shipping issues also pop up usually with regards to late deliveries or packages lost or not being able to be tracked.
Finally, complaints about prices also count here as customers describe how "the price tag is not worth the quality of the product" or simply that the product's price is too high.
Customer feedback in the Affogata platform is constantly being collected and analyzed. When companies want to learn and figure out, in real-time, what their customers think about their products and services, such analyzed data will help them understand what they are doing right and where their problem areas lie. And it all starts with the never-stopping AI tracking of the positive, the negatives, and the neutrals.
Sometimes sentiment can be subjective, especially with the most neutral types of conversations and Affogata lets people categorize easily each mention or feedback on the platform and change the sentiment according to their KPIs.