Do you have concerns about feedback analytics? We have the answers!
Are you someone who works daily with feedback, and would like to improve your team's efficiency with an automated solution? You likely have questions, reservations and concerns. We understand!
If you work with data as an analyst, then you probably already feel you spend too much time processing feedback data instead of helping your team act on it.
Painstakingly coding and analyzing user feedback in spreadsheets may feel like the most accurate way to extract insights from your data, but eventually, something’s got to give. It’s impossible to be objective when you’re doing the analysis by hand, not to mention the huge time commitment. Even with automated solutions, results can be delayed to the point where they’re useless - users move on, and you’ve lost the chance to do anything about the pain points and opportunities you’ve identified.
We understand if you have some concerns about how our platform can work for you, but we’re keen to put your fears to rest! Here are some of the biggest concerns our prospective customers bring to us, and how Thematic eliminates each and every one.
1. Will Thematic provide specific, actionable insights from my data?
We agree that applying a standard theme template to your data will miss necessary nuance.
With Thematic, the theme taxonomy is tailored to each of your datasets. Your feedback data is tagged with themes that are specific to your products and services, rather than industry-wide.
Thematic also provides the option to zoom in and out, going from general themes into sub-themes. You can get even more granular by viewing a summary and a list of specific issues, questions and requests linked to the original comments. By combining themes from text with quantitative data you can get even more specific insights. It literally doesn’t get more specific than Thematic!
2. Will you capture all the themes in a piece of feedback?
Every piece of feedback, and even sentences, can have multiple themes - users often mention several issues at once within the same piece of feedback. Thematic captures all available themes in each piece of feedback, so that you can get the full picture and analysis is truly valid.
This is a critical issue for all of our customers and has been a feature of Thematic from the start.
3. I have feedback in multiple languages. Will I need to translate it?
We use the Google Translate API to translate any non-English data into English prior to analysis.
A customer of ours is a large multi-national telco in Europe. They used to employ translators to analyze feedback from European languages into English. For the past 5 years, they've been a customer of Thematic. They have assessed the accuracy of themes based on translated versions of the text and are happy with the results. Similarly, Jetstar uses Thematic to analyze feedback in Japanese as well as English. This is a common requirement for our customers.
Why don’t we analyze feedback in its native language? Machine translation has become accurate enough for discovery of themes. Also, English is an easier language for NLP/AI to ‘crack’. Finally, by seeing all feedback analyzed in a single language you can compare insights across geographies and languages.
4. We currently use a hierarchy of topics. Can Thematic be trained to match that?
Thematic organizes themes into a taxonomy of base themes and sub-themes. It’s fully self-supervised and doesn’t require any training. The resulting taxonomy makes your data much easier to navigate. And it’s incredibly flexible and adaptable - you can merge, move, add or delete themes to refine your taxonomy exactly as you wish.
What our customer Atlassian loves about Thematic compared to other AI powered solutions is the specificity of themes and the ability to adjust the results easily. In Thematic you can make a change any time, revert to a prior version, and work with AI to refine and discover more themes.
5. How noisy can the data be - do I have to clean it first?
Good news! Thematic filters out non-feedback like greetings, irrelevant chatter, and mentions that aren’t feedback.
Not all data is equal. Depending on the source, feedback may be mixed up with non-feedback. In addition to details of user problems, support conversations contain discussion, standard and scripted responses, and greetings. Some solutions will expect you to provide data already cleaned of this additional chat - and data cleaning is tedious and time consuming. Thematic knows what’s feedback and what’s not, so you don’t have to specially prepare your datasets pre-analysis.
6. What's the advantage of working with a specialist solution, instead of an all-in-one option?
Most companies receive customer feedback through a wide range of sources, both internal (surveys, contact center, support tickets, complaints, CRM notes) and external (review sites, forums, social media).
If you limit yourself to just one tool, e.g. support tickets, you’ll be biased to those customers who use them. If you use the different analysis tools built into each solution, you have no consistency. The result? You lose the full picture of your user’s experience. Thematic is agnostic to how you collect feedback and is focused on aggregating this feedback for a holistic view and accurate analysis.
7. At what point will analysis become faster and more accurate (compared to me doing the analysis myself)?
An automated solution means consistent analysis, less bias in decision making and huge time savings. If you’re processing upwards of 1000 pieces of feedback per month then it's well worth the investment.
You can then focus on leveraging results, rather than being bogged down processing your feedback data!
8. How long does it take to set up?
Setup is possible in days, so you can pull accurate insights in near real time. We’ll help you implement, but you can also spend more time to fine-tune results if your use case requires it.
9. We have data scientists in house that can build a solution.
An in-house solution sounds like a dream, but building a feedback analysis solution is complex and difficult. While it’s easy to get useful results for a one-off experiment, most companies opt to buy instead of build to get a viable long-term solution. To build a solution in-house, you need a team of experienced data scientists, software engineers and data analysts. You’re looking at a lead time of 1-2 years. That’s a long time to wait for insights that you need now.
From making sure themes are meaningful to accounting for data privacy and GDPR, there are serious pitfalls to consider. Manual feedback analysis is painful, but it doesn’t compare to the frustration of building a solution that doesn’t work. We’ve broken down the steps to creating your own solution here.
Talk to your data scientists, and find out whether their time may be better spent leveraging an existing solution like Thematic. Thematic automates the coding of qualitative feedback regardless of source. There’s no need to set up themes or categories in advance. Simply upload your data or use our one-click integrations and wait for AI to do its magic. Your team can manually edit the codes to further refine their accuracy. Experiments conducted indicate that Thematic’s automated coding is just as accurate as manual coding.
10. I don’t want to get stuck paying for something that’s difficult to use.
Here’s what we’ve heard from our customers:
“I just recorded a 5-min loom about how to use Thematic and my team could get the insights they need themselves.”
“I was surprised how quickly I could start getting answers from Thematic, it’s really easy to use.”
We’re here to ensure you get the most value possible from Thematic. We have a dedicated customer success team, as well as data science experts available to help you. Your assigned customer success manager will guide you through onboarding, and have regular one-on-one meetings with you/your team going forward.
We have a customer slack channel, and a slack insights community that you can join. Our regularly updated public knowledge base has loads of how-to articles and videos, and we also have 24/7 in-product live chat with our support team.
11. How do I get data in and out of Thematic, and where is it stored?
We have several native integrations, and we'll work with you if you need to import data via an API. Your data remains yours - we don't hold any data hostage and you can extract results and visualizations in the way that works best for you.
Several of our customers pull tagged data into their data warehouses, and we’ll be happy to help make this happen for you.
By pulling tagged data into your data warehouse, you can combine feedback with other customer data that isn't uploaded to Thematic. You can create custom reports directly in your dashboard tools like Power BI, Looker or Tableau.
12. We already have a categorization framework, and we don’t want to lose our existing taxonomy.
We can help you align Thematic’s themes with your existing taxonomy. In Thematic you can have multiple theme structures or a single master-taxonomy. You can use them as different lenses to view feedback.
Some of our customers use a master-taxonomy of themes and then apply it to different datasets. Then they run theme discovery using built-in AI to get nuanced insights from each specific dataset. There are many ways that Thematic can help you refine and expand your existing taxonomy moving forward.
13. Will my results really be accurate?
Accuracy in feedback analytics is subjective. Even people disagree when deciding how to code a piece of feedback, and we know that Thematic’s automated coding is just as accurate as manual coding. To be thorough, our team uses an internal accuracy evaluation. We take a random sample of the data and a person manually verifies the tagging accuracy. An accuracy score is calculated at the end to ensure your results will be as accurate as possible.
A customer of ours once took a sample of 300 pieces of feedback, and reported that we were 95% accurate. The best thing about Thematic is the flexibility to keep iterating and see accuracy increase within minutes.
14. How do I know Thematic will be a good fit for us?
We’re not going to lie: we’re awesome, but there are some cases where we recommend another solution.
Thematic cannot natively process video or audio chat; it needs to be transcribed first. That being said, transcription services are improving all the time, and if you’re happy to connect transcribed data we’ll be able to help.
For surveys, Thematic requires at least 100 free-text responses to each open-ended question. If you have very small datasets we’re probably not the right choice for you.
We’re typically not a good fit for interview transcripts. This is because they tend to have a low number of participants and a wide variety of open-ended questions.
And last but not least: Thematic does not offer an API-only white-label solution for use within an existing CXM (customer experience management) or MR (market research) platform.
If you still have questions, we're more than happy to help - book a demo with our team, or download our free Buyer's Guide for an exhaustive list of things to consider when looking for an automated feedback analytics solution.