For decades, qualitative research has been fundamental to the success of companies of all sizes. However, choosing the right qualitative research transcription services could determine if your research project ends in success or failure.
Before analyzing any qualitative data, you must first convert the audio recording into a transcript. You can do so either by yourself, with an in-house team, outsourcing the work to an agency, using automated transcription software, or hiring a freelance transcriptionist.
With the multitude of options available in the market, how can you choose the best transcription service for your specific needs?
If you don’t know what questions to ask yourself when choosing a transcription service, don’t worry - we’re here to help.
In this in-depth guide, we list out FIVE important things to consider when choosing the right transcription service for your qualitative research project. In summary, they are:
Before we dive into what makes a reliable transcription service, let’s take a step back and briefly go through what transcription services are.
Transcription service providers take your audio or video recording, process it, and convert it into a transcript before sharing via an encrypted medium. There are two main types of transcription services: manual and automated transcription.
Trained transcriptionists listen to your recording, often several times, and convert it into written text through either verbatim or intelligent transcription.
Human transcriptions are also the most reliable way to obtain 99%+ accurate transcripts, making them more suitable for qualitative research analysis or medical transcriptions than automated transcription.
You upload your recording to a data transcription software, and it generates a mostly accurate transcript in mere minutes.
Automated transcriptions are suitable for quickly transcribing recordings and building a media database without breaking the bank. Once you have a decently sized archive, you can then sort out important data sets to be cleaned up by professional transcriptionists.
Read more: What is Data Transcription in Qualitative Research: Everything You Need to Know
Now that you have a general idea of the types of transcription services available, let’s look into five major criteria to determine the best transcription service for your qualitative research.
Human transcriptions are 99%+ accurate, whereas automatic transcriptions are usually 80-95% accurate.
However, while human transcriptions are more accurate, that does not necessarily mean that they are better than automated transcriptions - the best qualitative research transcription services are those that serve your specific needs.
Manual transcriptions are more suitable if you need high-quality and highly accurate data sets. These detailed data transcriptions are suitable for qualitative research, medical transcriptionist, recording legal proceedings, and transcribing recordings with background noise.
On the other hand, automated transcriptions are a great starting point for building a media database. Then, you can pick data sets that are important to you to clean up with human transcription services before conducting further qualitative analysis.
Automated transcriptions are also great if you need something optimized for general data structuring and searchability since the saved data sets are stored in one centralized system.
Professional transcriptionists can replay the recording several times and make any required adjustments. On the other hand, automated transcriptions are usually trained to recognize common words used in daily conversations. As a result, they may not recognize non-verbal cues, complex terminologies or identify different speakers.
Nonetheless, humans aren’t perfect either as demonstrated in this study that transcriptionists have an average of 4% word error rate (WER) when transcribing conversational speech. In contrast, automated speech recognition (ASR) with a 20% WER is considered acceptable but still not practical enough for qualitative research.
For example, here is a comparison between the original script and the converted transcript via automated transcription software.
As shown above, there is a deviation between the accuracy of the manual and automated transcription. The most distinct errors are punctuation usage, inability to recognize named entities and different sentences.
Whether you choose human or automated transcription services, there are four factors that affect their accuracies.
Technical and industry-specific language: Automated transcription softwares are mostly trained to recognize common, daily words. As a result, they may not identify technical jargon and other industry-specific terms. If you are hiring a professional transcriptionist, ensure that they must be familiar with the industry you’re in and its technical terms.
Audio quality: Both manual and automated transcription services are heavily dependent on the audio quality of the recording. Any background noise, cut-offs, lags, or static will affect the accuracy of the final transcript.
Regional differences: While English is a global language, it contains many nuances specific to different countries worldwide. Variations in dialects, accents, spelling, and grammar may be confusing if your transcriptionist is unfamiliar with those regional differences.
Automated qualitative research data transcription has a 1:1 turnaround time. In other words, it takes automated transcription tools approximately 1 hour to convert 1 hour of audio. Fortunately, a system like Speak runs the transcription and analysis in parallel for multiple files with the same turnaround time -- perfect for bulk transcription.
Let’s say you upload 10 files, each one hour long. Instead of a one-hour turnaround per file, you’ll have all files transcribed in one hour. In contrast, many automated transcription systems run in sequence - meaning they won’t transcribe your next file until the previous file is transcribed.
While it may not be as accurate as manual transcription (80-95% vs. 99% accuracy), automated data transcription is excellent for quick, cheaper turnaround for bulk transcription.
If the transcripts’ quality and accuracy are non-negotiable, you should opt for human transcription services instead. Usually, qualitative researchers hire expert transcriptionists because of their ability to provide almost 100% accurate data sets.
Depending on their skill level, transcriptionists may take 2-4 hours to convert 1 hour of audio into a transcript. To obtain the high accuracy requirement, they must replay the audio several times, ensure that every punctuation is correct, and format the transcript according to verbatim or intelligent transcription standards.
There are several options for manual transcription services. You can:
Quality and accuracy of transcription services go hand in hand, especially when hiring a transcriptionist. The transcriptionist should possess at least native-level English. If you’re hiring a vendor, it should preferably be an established transcription service provider who has been in the business for years.
Moreover, just because someone has native-level English does not necessarily mean that they are a qualified transcriptionist. As a guide, here are 5 qualifications that you should look for in any professional transcriber.
Good knowledge of the industry: Your chosen transcriber must be familiar with the common technical jargon in your industry. If the transcriptionist does not have industry experience, they should have versatile research skills to quickly get up to speed with the industry language.
Grammar and punctuation: Your chosen transcription service provider must know everything about grammar, punctuation, spelling, and all other rules in the English language.
Impeccable listening skills: Naturally, a transcriptionist must have excellent listening skills to convert speech into text efficiently. Your chosen transcriptionist should also use high-quality headphones and software to isolate the audio channels in the recording to hear conflicting dialogue clearly.
Training and certification: There is no educational barrier to entry for being a transcriptionist. However, the transcriber should possess at least a high-school diploma or, preferably, a relevant college degree. There are also online training resources available such as those at Transcription Certification Institute (TCI).
Ability to meet deadlines: Any person or company who provides qualitative research transcription services must have a proven track record of meeting deadlines. If your transcript arrives late, it will affect every subsequent step of the research process and result in a delayed or, worse, a shelved research project.
Before hiring any transcription service provider, do a background check and ensure that they have a reputable track record of managing sensitive information. Also, ask how they intend to secure the delivery of transcripts and own database systems from potential hackers.
Here is a checklist of confidentiality compliances that qualitative research transcription services could follow:
ISO 27001 spells out the requirements for an information security management system. While not obligatory, it highlights the capability of companies to manage financial information, intellectual property, and other sensitive information provided by their customers.
On the other hand, ISO 9001 is a customer-oriented certification that ensures companies provide consistent, high-quality products and services.
Any company or freelancer that provides qualitative research transcription services must sign a non-disclosure agreement (NDA). The NDA will ensure that all parties involved will be able to share sensitive information without it falling into the hands of the competitors.
HIPAA stands for the Health Insurance Portability and Accountability Act signed in 1996. Generally, HIPAA compliance sets the standard for sensitive patient data protection and is mandatory for transcription services dealing with medical transcriptions.
The General Data Protection Regulation (GDPR) is a fairly recent law that protects the privacy of every EU citizen. It also covers areas such as data processing, accountability for whoever controls the sensitive data, and data security. Any business that violates the GDPR will be fined a maximum of 4% of its total global revenue.
Encryption refers to securing digital data through the use of “keys,” hence its namesake. With social engineering and cyberhacking posing an ever-constant threat, encryption provides a layer of defense against any unwarranted data phishers.
Any website that provides qualitative research transcription services must be encrypted HTTPS. HTTPS encryption secures the connection between the website and your browser and prevents unwanted intruders from stealing sensitive data.
Furthermore, transcription services that involve transferring information through log-in portals must use SSL or TLS encryption. Put simply, SSL and TLS are security layers that go hand-in-hand with HTTPS encryption to protect sensitive data transfer between two systems.
Automated qualitative research transcription services are more cost-effective than human transcription services.
The average cost for automated transcriptions is $0.15 per audio minute, while professional transcriptions can cost $0.50 to $2.00 per audio minute.
Scalability: Some transcription software can convert multiple audio files simultaneously in the same amount of time it would take to convert one file. Moreover, human transcription services have large teams of transcriptionists and project managers to handle any fluctuations in workload.
Return on investment: For the highest ROI possible, you should choose a transcription service that is reliable and commits minimal errors. You also have to consider that while human transcription services are more expensive upfront, the accurate datasets they produce can lead to invaluable actionable insights.
Value-added services: Some transcription services provide supplementary services such as sentiment analysis, keyword mapping, and even free post-delivery corrections.
There are many human and automated transcription service providers available -- Speak Ai, Otter Ai, Rev, Descript, Monkeylearn, Temi, TranscribeMe, and more. All of these companies follow all if not most of the criteria established above.
If you are interested in learning more about Speak Ai and its alternatives:
Here at Speak Ai, our transcription services are simple yet highly effective in producing outstandingly accurate transcripts. We offer no-code solutions for individuals and organizations with the ability to manage all their research workflows in one centralized database.
Aside from manual and automated transcriptions, we allow you to take your qualitative research one step further with sentiment analysis tools and API integrations with Evernote, Zapier, and WordPress.
If you are keen on learning more about how Speak Ai can bring more value to your organization, get started with your 7-day trial or contact us at success@speakai.co.
There are five important aspects that you must consider when choosing the transcription service that best suits your qualitative research needs. They are:
Accuracy: The accuracy of human (99%+) and automated transcription (80%~) is affected by the audio quality, regional differences, and industry-specific language.
Turnaround time: Trained transcriptionists can convert 1 hour of audio in around 2 hours, whereas transcription softwares usually boasts a 1:1 turnaround time.
Quality: Transcriptionists must be familiar with the industry you’re in, have strong research and listening skills, preferably have training and certification, and have extensive knowledge about grammar and punctuation.
Confidentiality: The transcription service provider should be ISO 27001 and ISO 9001 compliant and must be HIPAA compliant when dealing with medical transcriptions. Moreover, data sharing between you and the vendor should be protected with HTTPS, SSL, and TLS encryptions.
Cost: Manual transcriptions cost higher but also provide more accurate transcripts than automated transcription. Also, consider each transcription service’s complementary services, such as sentiment analysis and named entity recognition.
Get a 7-day fully-featured trial.
Powered by Speak Ai Inc. Made in Canada with 
Use Speak's powerful AI to transcribe, analyze, automate and produce incredible insights for you and your team.
