Most will also be aware there is a considerable resource in customer feedback as it can be an unfiltered way of understanding what customers actually think, as well as what they need and want. But again, this is hard to do. Feedback spreads across different platforms, from call centres to online reviews, internet chat sessions, and even in-store. Connecting different platforms of collection methods is possible but, it’s too easy to create gaps where insight can be lost. Within many organisations customer feedback can be seen as a PR issue or a problem that needs to be dealt with. Read on, and we’ll show you how it can be something you could actively encourage.A few seconds looking on the internet for methods of easing your customer feedback analysis process can be overwhelming. There is a massive arsenal of tools available, but they often overlap or are unclear about what they can’t do. In many cases, enterprises are still processing feedback manually, which can be both inaccurate and inconsistent.
The result is a mass of data which is enormous and difficult to decipher. Like most business problems, this can be broken down into manageable chunks, but even then, the complexity behind the fragments can be technically challenging.
Why bother being customer-centric?
According to research carried out by Deloitte and Touche, customer-centric companies are 60% more profitable than those not focused on the customer. This proves that the moment you put the customer at the core of your business, you’re already one step closer to success.
Even though most businesses claim to be “customer first”, they aren’t taking steps actually to be there for their customers. An incredible 42% don’t collect customer data at all.
Customer feedback is useful for product development and marketing- it can be the quickest way of finding out what people like and don’t like about your product or service. But it is also part of a virtuous circle. For a start, being able to deal with common customer issues lowers service costs. Further, understanding and responding to feedback has a demonstrable impact on loyalty, which itself reduces customer acquisition costs. Most large enterprises will have some method or proactively collecting reviews and customer thoughts during or after a purchase. But is this enough?
All feedback is valuable, but especially unsolicited feedback, the feedback you don’t ask for. This, of course, makes life hard. Firstly because it is usually unstructured so won’t fit in your usual categorization. Secondly, because unsolicited feedback is usually written by motivated customers who, by nature,can be either be over-positive or over-negative. This is usually called the ‘J-Curve’, where what seems to be an average 4-star review has a large number of motivated customers giving one star. However, it is through analysis of this group of motivated low-scoring customers where you may find gold!
See how the world’s most customer-centric brands are using Wonderflow on our Use Case page.
This leads you to one of the major outcomes of being customer focussed- accepting not only that all feedback is valuable and should be acknowledged but that part of the strategy could actually be to encourage more feedback. But we can only do this once we are capable of dealing with the feedback we have.
How do we collect and analyze customer feedback?
Analysis of feedback is a multi-step process and it’s essential to see the complexity and decisions to be made with each. With each stage, it is critical to know what the right inputs and outputs are, which means knowing the right question to ask.
First of all, we have to collect all the raw data. Don’t be fooled into thinking this is a trivial task. Some of the data will be in-house like call centre logs, emails to CS, NPS surveys, or existing customer feedback programs. Some will be external, like indirect retailers or review websites, or commissioned market research. External data also means that a large proportion of the data is public too. A lot of organisations still use web-scraping apps, some open source, which can do this quickly. Since the number of channels for a global firm can be enormous, immediately, we see how what seems a simple task of collection has begun to grow! The input may be massive , but the output here has to be simple: you need to have your raw data in one place.
Data preparation is where we get into the meat of the problem. Information collected will have duplicates, irrelevant information or ‘blank’ data which needs to be cleaned out. There will invariably be typos. Data from public sources may require header and footer data to be scraped as well. All information will be anonymized. Some will be in different languages, and even if you collect from the same language in one country, there will be different idioms and figures of speech. And that’s before we get to dialects, the vast numbers of spoken and written variations of French, Arabic, German, even English dialects globally. To get to this level requires some sophistication, and either intelligent software or human text scanning.
At this stage, trade-offs will be needed: time, accuracy, consistency, volume of data. You can read more about the pros and cons of human analysis here. This leads to one of the first major decision points: what is the scope of the analysis you want to start? How local do you want or need to be? The endpoint of this stage is to get to usable organised data collected with meaningful scope.
Once you have the data, the information needs to be coded to fit your business requirements. When customers are leaving reviews online, they are not filling your pre-set formats. Actual issues can be buried in the text and may not be written about directly. This requires an understanding of subtlety within the language.It’s possible to run this process manually and then add on a text analysis solution. This alone can work, but by its nature is a shallow solution- it will only look for what you ask it to. Wonderflow uses Natural Language Processing. NLP is a branch of artificial intelligence that helps computers to derive meaning from human language. This sounds highly sophisticated, but you’ve almost certainly used it before- NLP is how search engines search predicts your query, and your email provides sorts your inbox. Here’s an article with loads of great examples. NLP has a number of different aspects. For example:
- NLP allows for precise sentiment analysis. Simply put, this is a structured way of understanding how people feel through the language they use. For example, using the character ‘s’ to signify the plurality of items, or expressions such as ‘fantastic’ or ‘disappointing’ to sense how they feel, or considering the variation of opinion in between positive and negative. Even stronger is analysis about particular product aspects or features and gauging the relevance of those features to the overall customer satisfaction. A time aspect can be introduced to see the effect of a particular product or service changes. There’s more about sentiment analysis here including how it can be used to detect complicated writing such as sarcasm.
- Data Enrichment: At this stage, there are numerous ways that the current dataset can be enriched. For example, descriptive analytics takes the results so far and prepares a “summary view” of facts and figures in an understandable format. This can be used to either provide knowledge immediately or prepare data for further analysis. It can pull together customer feedback and analyze it, showing how frequently different pros and cons are mentioned. Even at this stage, this gives management a solid overview of a product’s best qualities. Read more about this here.
- Another hot topic is predictive analytics. Predictive analytic solutions forecast future outcomes by reading and interpreting historical data. Through the use of statistics, advanced algorithms, and machine learning, quantitative and qualitative information is transformed into predictions. Predictive analytics can he highly sophisticated and of course, that means it can be highly complex as well. Depending on your industry, internal users and data sources, it might not be right for you. This link will help you understand if predictive analytics is right for your enterprise.
Analysis and NLP are at the heart of what Wonderflow does. Our NLP analysis solution aims to simplify the process and remove the technological complexity dramatically. But even then, the smartest analysis needs to be used and presented within the organisation.
Analyzing multi-language customer feedback in large volume from different sources is a complex process. With an AI-based technology and years of experience, Wonderflow is helping global brands to become customer-centric. Find out more about our solution.
You will now have a huge set of data and analysis about history or even the future of your customers. Somehow this will need to be shared with both internal and external stakeholders in your organisation. In almost every enterprise setting, this means having reporting tools which allow for APIs and connectors. You will already have established tested reporting systems and dashboards of KPIs.
At this point, you will need an organizational understanding of how insights will be used. Often customer feedback is limited to just marketing teams, in which case you will focus on KPIs like NPS or CES. At the other extreme, the results can be available across the organization so anybody can query them. This requires either connecting the results and data to another dashboard so they can be queried instantly or processing all the KPIs yourself, so users see only the result numbers.
Again, there are trade-offs between processing time and capability and more to be considered. Do you need real-time data visualisation or customisable dashboards and charts? Will users have to quickly and efficiently digest large volumes of data? Will the results go in a single quarterly report are will they be accessible continually?
This need for flexibility in reporting and actionable insight drives our reporting tool the Wonderboard. Whatever you choose, Wonderflow’s solution offers unlimited seats.
The final step in working with feedback is to develop your hard work into insights- knowledge you can actually use! A 2017 report from Temkin shows that less than 25% of companies consider themselves good at making changes to the business based on the insights.
Not all insights are actionable. Actionable insights are not more information or more data. To point out the seemingly obvious: insights, information and raw data are not one and the same. Part of the solution has been addressed in previous steps- using structured and unstructured data, internal and external sources, ensuring you are asking the right questions. The key to building insights is taking results and ensuring they are linked to business processes- you want insight that can actually make a change- it does not confirm what you already know.So what is an insight? An insight is a finding that contradicts your knowledge, confirms or denies your suspicions, or quantifies the importance- only you can figure this out within the context of what’s essential to your business. The starting point here is to understand what your team is being asked to do and why. You can read more about using insights here.
Conclusion: the way forward
It’s essential to build a feedback loop that works in today’s market. Consumers are expecting to be heard,which means you have to listen, understand and act! Spotting trends, understanding market responses in specific geographies or user groups are key to driving sales and product success. If you are already looking at customer feedback analysis, ask yourself and your peers if you are making the most out of it. Are you at a stage where you are just “managing” feedback or treating it as a problem? Where will the results of your analysis go? Is it product teams? Marketing? Or organization-wide?
There is no question that it can be a complex issue and that many organisations have embedded analytics solutions that have evolved from different vendors over time.
However, understanding customer feedback and making use of it is the cornerstone of developing a customer-centric strategy. It can be predictive and powerful and create a genuine voice for the consumer within the organisation.