Month: October 2020

If you’re dictating your blog posts, novels, books, articles, and any other type of written content, you have to turn your audio into text. As I’ve covered in other posts in this blog, you have several options available to you.

In this post, I’d like to focus on dictation bloggers and writers who want to hire manual transcribers. Generally speaking, this is called intelligent verbatim transcription.

You’re not asking them to completely rework or reorganize your output. You’re not asking them to make stylistic or editorial substantive changes. Instead, you just want them to weed out your “uhh” and “ahh” and obvious grammatical mistakes.

You also want them to be intelligent enough to figure out when you’re editing yourself or simply reiterating a passage that you just dictated. Make no mistake, if you are looking for the highest quality transcription, without tripping yourself up in terms of dictation speed, intelligent verbatim transcription is the way to go.

This is my personal preference. I understand that other writers would rather dictate into their mobile phone straight into a speech to text transcription app. I covered that in another post as well. I even listed out the top 20 voice recorder apps that you could use.

You could then plug those apps into automated transcription. Here’s a little bit of warning though. As somebody who has tried digital transcription, I can safely say that that method of turning speech to text has a long way to go. It definitely needs work.

I’m dictating this blog post right now and my transcriber is intelligent enough to know what the natural breaks in the dictation is. My transcriber can also detect any emphasis that I may be placing on certain words.

This goes a long way in setting up effective sentence construction. Also, I don’t have to worry about pausing and saying “exclamation mark” and “period” at the end of every sentence because my transcriber knows when to naturally put punctuation.

In fact, by just listening to my tone of voice, they would know the difference between a period, a question mark, or, in rare cases, an exclamation point. Not so with automated transcription technology.

They do have their place, but if you are dictating a 700 to 900 page epic novel like I did, intelligent verbatim transcription is the way to go. With that out of the way, here are the top 30 most common manual transcription mistakes transcribers make.

Whether you are hiring a transcriber to do intelligent verbatim transcription or some other type of speech to text work, pay close attention to how they handle these mistakes.

If you see them committing the same errors over and over, you might want to ask their project manager to tighten up the quality. But if the issues continue, you might want to switch to another service.

The 30 most common manual transcription mistakes

Mistake #1: Inaccurate transcription

Let’s face it, the English language is very tricky. At first blush, when you say these words very quickly, they seem to be the same. You have to read the context to get the right spelling. I can say “seem” and then “seam”.

Since you’re reading this, you know these are 2 totally different words applied to completely different contexts. But when you’re listening to speech audio, they sound the same. The same goes with having “seen” something and arriving at a “scene”.

If you’re reading these sentences, the correct spelling jumps out at you. But if you’re just listening to it and you’re not paying attention, it’s easy to mistake “seen” with “scene”. I wish these were the only 2 words that sound very similar in English. But there are tons of them.

This is why it’s really important to look at your transcript, no matter how quickly, once you get it back from your transcriber.

Mistake #2: Bad basic grammar

As much as possible, you should outsource your transcription work to people who have a good command of the English language. By doing so, you can rest assured that they at least have a working knowledge of proper capitalization, spelling, and punctuation.

Sadly, even an experienced transcriber can still screw this up. Again, you have to go through an extra quality control step. Maybe you should tell them to plug their work into online tools like Grammarly or some other free online grammar checker.

When it comes to misspelling, this is a very easy problem to spot. But punctuation can be quite tricky especially if the person is a complete newbie. Capitalization rules are also pretty straightforward, but again, people do make mistakes especially if they’re rushing through a transcription.

One simple trick that I use is to plug in the transcribed text into Google Docs. Google Docs uses artificial intelligence to detect errors. When you see a blue underline under a text and it’s not hyperlinked, there’s something wrong with that text.

Click on it and you will see Google Doc’s recommended grammatical correction. Sadly, not all its recommendations are 100% accurate. There is such a thing as a false positive grammar detection error with Google Docs.

This is where you use your basic English skills to see if the recommendation makes sense or not. When it comes to misspellings, on the other hand, Google Docs is pretty straightforward. If you see a word with a red line underneath it, it is spelled wrong.

9 times out of 10, you can just go with the recommended change and it would be absolutely correct.

Mistake #3: Paraphrasing or rearranging the speaker’s words

This is a big no no. As much as possible, manual transcribers who are doing intelligent verbatim transcriptions must stick to the text as dictated by the speaker. In other words, they have to follow the form and the order of the text being dictated.

If they were to take portions of sentences and mix them around, it’s going to throw the whole paragraph or sentence out of whack. 9 times out of 10, it’s not going to make any sense. Sometimes, I do instruct my transcriber to rearrange portions when I know that I missed a step.

For example, when I’m dictating product reviews, sometimes I overlook key benefits to the customer and I have to go back. Instead of dictating the whole thing, I would say to the transcriber “Please take this portion” and I would specify which section it is, “and plug it into a section after a subheading.”

Generally speaking, your transcriber would know what to do and take care of the issue. But that’s probably the only situation where rearranging or paraphrasing would make sense. Otherwise, they’re not doing a good job if they just completely override your judgement.

You’re the writer. They’re the transcriber. These are 2 totally different roles.

Mistake #4: The transcriber adds additional information

If you’re a blogger getting paid by publishers on a per word basis, you might think this is a good thing. After all, when it comes to writing for money, the more, the merrier, right? Not necessarily.

When the transcriber makes it a habit to almost always add additional information, this makes your job harder. You’re going to have to drop whatever you’re doing, read the transcription from beginning to end, and make sure that whatever was added actually added value to the transcription.

Unfortunately, this doesn’t always happen. If anything, it adds confusion or it takes your train of thought in the text to a different direction. That’s not always a good thing. To minimize any of that unnecessary hassles, insist on your transcription service provider to focus on what’s in the audio.

There’s no need to add additional information. If, for example, you’re dictating a legal brief for an attorney client, it’s on you if you missed a key fact and the transcriber is under no obligation to step in and try to plug that hole.

The same goes with a novel. I know, I’ve made this mistake several times before. I’m in page 500 trying to flesh out the character and bring color to the conflict, but I would overlook a very important detail that the transcriber has already covered and transcribed on page 200.

The person should just transcribe what I dictate. They know full well that I missed a very important fork, maybe it’s logical, maybe it’s an emotional development, maybe it’s some sort of psychological revelation that can prove crucial to the story line, but that’s on me.

I dropped the ball and I have to deal with the circumstances. This is much better than just giving your transcriber a blank check to plug in additional information. First of all, they’re not getting paid enough to do that because that’s high level editorial stuff, and second, it would probably trigger all sorts of issues with the final work.

So it’s much better to just set clear rules and boundaries with your transcriber regarding additional information.

Mistake #5: Mislabeling speakers

As I’ve written in this blog before, one of the fastest and most powerful ways you can create free content is through recording interviews. That’s right. Just go to all the conventions that you can in your field and interview experts.

A lot of these experts would love to talk their heads off because it shows how much they know and they are promoting their brand by talking to you. It doesn’t matter how small your outlet or audience may be, the more they drill down into that specialty niche market audience, the stronger their brand becomes.

It’s a win-win situation. The problem is when you are interviewing different people at the same time, the transcriber might mislabel the speakers. Believe it or not, a lot of people have similar sounding voices. You have to really lean closely and pay close attention to the audio.

As the person who recorded this material, you wouldn’t have a problem because you know who’s who. They were in front of you when they were talking. Maybe they’re talking on top of each other, but you know who was who.

Good luck trying to figure that out if you’re the transcriber. It’s really important to make sure that you deal with a company that filters or trains transcribers so they can accurately label speakers. This goes a long way in making sure you collect all the streams of content from many different sources during a group interview.

You can then have these transcribed and you can filter these out yourself and have it edited later and you have yourself a nice stack of content.

Mistake #6: Time stamp inaccuracies

This usually applies to transcriptions of meetings and multiple speakers. This also applies to transcriptions that are either going to be used in court or crossed referenced.

This is usually not that big of a deal when it comes to transcribing blog posts, articles, novels, and other creative works. But this can be an issue especially if you need to refer back to the audio and correlate it with the transcribed text.

If you can’t do that, that’s going to be a problem.

Edited transcription mistakes

In the following series of errors, I’m going to cover edited transcriptions. This is a very different type of service compared to intelligent verbatim transcriptions.

Edited transcriptions require the transcriber or editor of the transcript to make some judgment calls as to how to condense ideas or to reword certain concepts throughout the transcription so it makes better sense.

For the most part, this doesn’t really apply to dictation blogging because depending on how aggressive the editor is, a lot of your personality as well as your unique point of view can eithter be watered down or completely cut out.

But there is definitely a market for edited transcription in other industries, but not necessarily in creative work. Still, I’m going to cover it here because these types of mistakes are quite common and should be detected and fixed.

This is especially true if you don’t really care much about putting your personal stamp in your output. Maybe you just want to get as much content out there, so you hire companies that do edited transcriptions.

Be on the lookout for these problems.

Mistake #7: The general idea of the text isn’t properly maintained throughout the whole transcription

This usually happens when a significant portion of the dictation is reworded or cut up and placed in other parts. What happens then is what should’ve been a point of clarification that kind of recaps or reiterates what the whole piece was about is missing.

This is not a problem if the transcribed piece is, let’s say 1,000 words. Usually, people can follow an idea close enough to maintain some sort of consistency of meaning even though there’s some confusing passages.

But this issue can definitely blow up if you’re talking about a very big piece of text like a novel or a novella or a very long blog post. This is why it’s really important to make sure that key sentences are not only left in, but also emphasized through bold font or some sort of formatting device.

Mistake #8: Transcribing unfiltered verbatim

The main reason why you’re hiring an edited transcription company is for them to provide you with a very polished product. Many eyeballs have gone through the text. They have filtered out problematic parts and restructured them to preserve your idea.

In fact, if done well, edited transcription can produce an end product that is much better than you originally visualized. That does happen. You know there’s a problem when the transcriber includes gaps in the conversation and they transcribe mumbling or “uhh”, “ahh”, and other gibberish.

Usually, this is not acceptable. But there is space for this when you are dictating dialogue. I’ve done this before where there are several awkward moments between characters.

By keeping such verbal “tics” in the dialogue, you play up the conflict and tension between the characters. This can be effective, but generally speaking, in terms of edited transcription, it doesn’t make sense and it degrades the quality of the transcription.

Mistake #9: Failing to match the tone of the dictation

If you’re getting transcription work done in a foreign country where a lot of people speak English as a second language, you have to be careful about tone. A lot of ESL speakers have great English skills. The challenge is when English is taught overseas, it’s usually not taught in the colloquial context.

Instead, they learn it in an academic setting. So don’t be all that surprised when the transcription comes back and parts of it reads like a term paper. This is okay if your dictation is formal in tone. Maybe you’re talking about scientific research or you’re dictating a term paper for a client.

But if you are dictating a blog post on a review of stereo electronics, that’s going to be a problem. Generally, when people read reviews or other forms of consumer content, they get turned off by overly formalistic tones.

Make it a point to hire a transcriber or transcription agency who are staffed by people who can tell just by the way you dictate the overall tone that you’re looking for. Usually, if you start speaking in slang or using colloquialisms, they will pick up that this is an informal type of transcription. But you can’t be too sure.

Tone is important. Look at the transcription and make sure that it fits the tone that your audience is used to.

Mistake #10: Stammering, false starts, and gaps are included in the transcription

The reason why you’ve hired an edited transcription service company is that you’re looking for something that is smooth, flawless, and ready for publication. You might also have hired them for their editorial expertise.

Nothing slaps you in the face and reminds you that you made a bad decision than looking at the transcript and discovering that there are gaps. What’s worse is when the transcriber doesn’t include the gaps.

Basically, I instruct my transcribers to hit the shift and the key next to the zero on the keyboard to indicate a gap. Maybe these are missing words or sounds that they can’t quite figure out. However they do it, they let me know that there’s something that went wrong in the transcription.

I’ve been lucky enough to find really intelligent and professional transcribers that I don’t have to do this. But if you hire a company or a freelancer that just transcribes text and doesn’t let you know that there are issues, this is going to be a problem.

It’s a big headache actually cause you’re going to have to read the text at a surface level and then read it again at a deeper level and even then, you might not even know that there’s something wrong. You find out too late when a customer complains or your publisher gets you on the phone and calls you out. It’s a mess!

It’s really important to cover these rules first with your transcription provider. The good news is a lot of professional transcribers already know this. In fact, their consensus enough to email you and say “There were some rough spots here. Maybe your app or software tripped up during that segment. I’m just going to mark it in the transcription.”

That’s professional courtesy and most transcribers, and I know all of the people who work for me, are professional enough to do this. But just in case, you need to know that this is an issue. There are transcribers and transcription companies out there that don’t really care.

So they just crank out the text and they leave you to figure it out. The worst thing that you could do is to take that transcription and just post it automatically to your blog. That’s a recipe for disaster.

Mistake #11: Grammatical errors left untouched

As I mentioned above, I prefer that my transcribers plug in the transcribed text into Google Docs. They protect themselves when they do that. Not only does Google Docs have a built in grammar checker, but they use a cutting edge technology called machine learning.

What happens is there’s so many people all over the world using Google Doc and the stock grammar correction system would notify people. But here’s what’s awesome about Google. It actually pays attention to how people correct their grammar.

They don’t just store this locally. Google Docs learns from all these millions of corrections so that when other people make those mistakes in the future,Google Docs will recommend the best solution as judged by how most people fix that error.

The best part is Google Docs is absolutely free and it works very quickly. You just need to look for blue underlines and see if the proposed correction makes sense. This is why it’s inexcusable for edited transcripts to contain any kind of grammatical errors.

There is one exception though. If you are dictating a novel and it includes dialogue with people with bad grammar, this should remain. In that situation, the challenge is different. Your problem is to keep the grammatically incorrect portions in those key parts of your book or story.

Common errors made during verbatim transcription

There’s a big difference between edited transcriptions and verbatim transcriptions. Edited transcriptions give the transcriber a lot of leeway as to how to fix the text.

Verbatim transcription, on the other hand, comes in 2 flavors: Intelligent verbatim transcription and raw verbatim. As the term implies, raw verbatim transcription is just listening to the text, as spoken, and just typing whatever you hear.

If the person says “umm” or “ahh”, that makes it to the transcript. It’s raw. No refining, no editing, nothing! Word for word. In fact, some services even use sound for sound. So if somebody coughs, that makes it to the record.

Intelligent verbatim transcription is the middle way between edited transcription and raw verbatim transcription. In intelligent transcription, a lot of the “um”, stammers, false starts, and gaps are taken out, but the speaker is notified of any gaps or incomprehensible sections.

Also, the intelligent transcriber would pay attention to self-editing. This is, as far as I’m concerned, the biggest difference between raw and intelligent verbatim transcription. The way I dictate my blog posts, articles, novels, video scripts, and what have you is that I would speak out the general idea and, from time to time, I would correct myself.

Maybe I forgot something in a verbal list. Maybe I realized that there was a better way of phrasing something. When that happens, I would backtrack and it would seem that I’m saying the exact same sentence.

But in reality, there is a change in the sentence. I’m lucky enough to have intelligent transcribers who can detect this, so they use the last section. In a raw verbatim transcription, you don’t get that. You just get a raw dump of everything that you say.

So if you’re editing yourself several times within the same paragraph or, if you’re really unlucky, the same sentence, it would seem like you’re a broken record beause you keep saying the same sentence over and over again. It gets annoying because somebody’s gotta edit that and that person is usually you.

Keep this in mind as we go through the common verbatim transcription errors listed below.

Mistake #12: Gaps, stammers, false starts are taken out

For a variety of reasons, certain types of transcriptions need stammers, false starts, and gap fillers like “umm” and “uhh”. Usually, these are court transcripts. Generally, the reason why all this “junk” needs to remain in the record is to establish a context. It also gives the person reviewing the record some indications of credibility.

If, for example, somebody died and all the information they can get from that person takes the form of an audio recording, they would have to piece together the overall picture of that person’s trustworthiness and accuracy using audio.

In a court setting, they would need the transcription. By including these gap fillers, false starts, and verbal tics and whatnot, the decision makers and lawyers involved can rest assured that this is an accurate record of what the person actually said.

From this information, they can then make conclusions as to who’s trustworthy and who can easily be discounted. But generally speaking, if you are writing a novel, a blog post, or an article, these have to be taken out.

That’s why I suggest that your default standard when it comes to transcription is intelligent verbatim transcription. IVT, for short, doesn’t just knock out these items, but they also involve intelligent self-editing.

So if you’re repeating the same sentence over and over with slight changes, an intelligent transcriber will take the last version because that’s basically your statement to yourself and to them that “This is the sentence that I am most comfortable with.”

Mistake #13: “Misheard” words

I wish I could tell you that the only way to mishear something takes the form of similar sounding words like “seen” and “scene”. Unfortunately, I can’t say that. There are other words that have silent words or, when spoken really quickly, kind of morph into other words.

For example, “I’m going to go to the store.” You can transcribe that as “I’m gonna” or “I’m going to”. Verbatim transcription must stick to the words as pronounced. So it’s really important to look at accuracy errors because if you mishear certain words and you step in with what your judgement of what those should’ve been, you probably end up changing the meaning.

Transcription isn’t as good as it could be. In fact in some cases, it can just change the whole meaning altogether. The value of the transcription goes down the toilet. The company that I work for has hired transcribers before that has this issue.

They would then have to take the same audio file and reassign it to somebody who actually knows what they’re doing. It does happen and the problem is it kills a lot of production time.

Mistake #14: Shifts in breathing and ways of talking are omitted

Again, the whole point of raw verbatim transcription is to give you a written snapshot of the person giving the statement. If a criminal confession is transcribed, you best believe that the lawyers, both prosecution and defense, as well as the judge would want every single detail transcribed.

They pay attention to background noises, the overall tone, emotion, and breathing. Emotion can be indicated in the text. Make no mistake, there’s a big difference between a period, a comma, and an exclamation mark when it comes to emotion.

The person is doing a lousy job when they cut all this out and they just stick to the text when doing a raw verbatim transcription.

Mistake #15: Correcting grammatical errors

Again, with raw verbatim transcriptions, you want a textual snapshot of the audio so all the warts, wrinkles, lines, creases, and pimples of the speech have to be there. So if the person is making grammatical errors, leave it in because this is going to be used in court. This text has to accurately reflect the personhood of that individual.

Generally speaking, this type of raw verbatim transcript is only important when you can no longer talk to the speaker. Maybe they escaped jail, so they’re presently unavailable, or they died. But this type of transcript isn’t going to be that big of a deal if the prosecutor can directly cross examine or talk to the person who gave the statement.

But if all we have is the person’s audio record, the raw verbatim transcription must provide an unvarnished and authentic true to life depiction of that person. This includes grammar issues, verbal tics, and whatnot.

Mistake #16: Inability to transcribe every word

When you’re listening to a raw verbatim transcription, strap in. It’s going to be a long read because you’re going to go through every word that came out of the speaker’s mouth.

In intelligent verbal transcriptions, you don’t have to suffer through the “umm”s and “aah”s. You don’t have to hassle with the many times they have to repeat a sentence because they’re correcting themselves. You get all of that with raw verbatim transcriptions.

Mistake #17: Failure to show overlapping conversations

There is one type of extremely challenging raw verbatim transcription. Generally speaking, if you are transcribing a statement made by one person, it wouldn’t be an issue because this person would stop and then start. Maybe they have all sorts of weird verbal habits, but you can manage as a transcriber.

Things get really hairy when you have more than one speaker. You know what’s going to happen. They’re going to talk to each other. When things get heated, they try to interrupt each other. In many cases, more than one person is talking at the same time. Tempers might flare, so there’s a lot of cursing and emotional heat.

When you’re doing the raw verbatim transcription of all of this, you’re going to have to capture all of that. Basically, it’s hard enough to transcribe one person. Can you imagine transcribing several conversations at once?

To make things even worse, you have to tag the speaker as they talk to each other. If they’re Hollywood voiceover actors with really distinctive vocal sounds, this probably is not going to be a problem. But if they’re regular people off the street, there’s a good chance that some of them sound similar enough to each other.

In fact, people might not sound all that similar at first when they’re speaking in their normal voice. But you can bet that if they get all emotional, their voice can change. That’s when it can get extra tricky.

Mistake #18. Failure to indicate if the audio had pauses or failure to indicate how long these pauses are

Again, with verbatim transcriptions, you want a snapshot of what it’s like to actually talk to the person who is being interviewed.

And the person inspecting the transcript makes all sorts of findings based on how the person is talking and what kind of words were being used. They’re trying to paint a picture as to whether this person is believable or not.

This type of raw verbatim transcription is intended for fact-finders. Maybe this will happen in the context of a lawsuit, a criminal investigation, or maybe even a job interview.

The key is to get an accurate portrayal of the person speaking so the decision-maker or group of people making the decision can determine trustworthiness, credibility, or level of authority.

Unfortunately, when the transcriber leaves out pauses, this can get in the way of getting a realistic and accurate snapshot the decision-makers need.

Mistake #19. Failure to Note Unintelligible Sections of the Audio or Failure to Suggest Possible Transcription

As a general rule on raw verbatim transcription, transcribers cannot step in and place their educated guess as to what the missing word is. The best they can do is to put an underline and put question mark notations.

This should notify the reader that there’s something wrong with that part of the audio. They have to either listen to the audio again or look at the totality of what’s been said before and after to come up with some sort of idea of what the missing word is.

All of this goes out of the window if the transcriber completely skips the unintelligible part. They don’t put any sort of notation such as an underline or a question mark.

The speaker is saying something, and all of a sudden, there is a twist in the statement because there is a missing section.

Of course, the more words that are unintelligible, the more damage could be possibly done in raw verbatim transcriptions. In fact, this can be a fatal flaw.

It may well turn out that a few unintelligible parts here and there at the right points in the transcript may be enough to either throw out a testimony or a narration of an interviewee or weaken it so much that whatever is being decided can swing the other way.

Mistake #20. Spelling, Punctuation, and Formatting Are Completely Neglected

The funny thing about raw verbatim transcriptions is that a lot of people assume that when you record somebody talking that all the basic rules of English get thrown out. This is not true, not by a long shot.

If I start with just saying random words together with no relationship between them and completely forgetting about grammar rules, a transcriber will probably be able to turn that speech into text. I have no doubt about that.

But it’s not going to make much sense to you reading that stuff and it’s not going to make much sense as far as I’m concerned because, last time I checked, when people talk they usually do this to communicate, send a message, and understand each other.

They want to get something done or explained. There is a reason for talking and we need to remember this. That’s why most people who speak English or any other language follow grammar rules.

It doesn’t matter how deep a person’s vocabulary is or how many years of education they have. If you want to be an effective communicator, you have to follow basic grammar rules. These include sentence construction, punctuation, and even a certain flow which can be translated into a format.

You know you’re looking at a really bad piece of raw verbatim transcription when the transcriber thinks that they can disregard all rules because they are doing that type of transcription.

They think that as long as they hear certain words and transcribe it in a rough order, it is “authentic” enough. Absolutely wrong. If you want to stick to what the speaker is saying, you also have to follow the rules that they’re following.

This is the difference between a high-quality raw verbatim transcription and somebody who is just winging it. Again, important decisions might depend on what gets transcribed and what gets left out.

And just because you’re transcribing somebody who might not have an education or speaks very roughly, it doesn’t mean that you can assume that that person has no command of grammar.

Making this assumption is wrong because if that person is, in any way, functioning today in society that means that the person can communicate well enough. The transcription must at least reflect that fact.

Common Mistakes In Intelligent Verbatim Transcription

Intelligent verbatim transcriptions transcribe what the speaker is saying word for word but with several modifications.

First, you’re not going to transcribe “ums”, “ahs”, stammers, and gaps. As a transcriber, you will automatically clear that out because it doesn’t add value to the text.

Second, intelligent transcriptions detect whether the person speaking is editing himself or herself. As I mentioned earlier in this blog post, sometimes I say the same sentence three times when I dictate.

It’s not the exact same sentence. There are always changes either in the middle or near the end of the sentence. What I’m doing is spitting out different versions of the same sentence to try to communicate an idea.

The final version of the sentence is the one that I’m most happy with. It is the version that the transcriber should commit to writing.

Intelligent verbatim transcription uses this rule. You don’t blindly transcribe every single word coming out of the speaker’s mouth because if you transcribe everything along with the “ums” and the “ahs” and the gaps, it’s hard to make sense of all of that.

Especially if you are dictating a blog post, a novel, a book, or anything that is intended to be consumed as a ready-made form of content.

This is very different from transcribing every single word and sound where you are trying to create a snapshot of the speaker in the context of a murder case.

You best believe that both the prosecution and the defense would want a faithful transcription of the recorded testimony of the original witness who had died.

This way they can attack the person’s credibility or support it. When it comes to intelligent transcription, you’re not dealing with those considerations.

Instead, you want clear, easy-to-understand text that is faithful to the ideas being shared by the speaker. Keep this in mind when looking at the following errors:

Mistake #21. Including gap fillers

You know you’re dealing with lousy intelligent transcription when the transcriber puts in gap fillers like “um”, “ah”, and stammers. Their job after all is to edit that stuff out and get to the good stuff.

There is one exception though. It makes sense to keep this stuff in if I’m dictating conversations or dialogues. For example, if I’m dictating a novel, I would like to include “ums” and “ahs” to lend some authenticity to the exchange.

Usually, people stammer when they are under a lot of pressure or stress. The dialogue would seem more believable if that was left in. This is the only exception to the rule that I can see when it comes to this error.

Mistake #22. Excessive deletion of passages

You know you’re dealing with a newbie transcriber when they cut out whole paragraphs and shrink sentences. It’s one thing to have a good idea of what you think clear and effective writing should look like. It’s another to override the speaker.

Please understand that it’s the speaker who is the author of the transcription. The transcriber is simply the means or the intermediary that turns the speech into text.

They have a certain level of discretion but it doesn’t go so far as cutting out important details, automatically shortening sentences, or otherwise making very important strategic editing decisions without the previous approval or instruction of the speaker.

If you hire a transcription company that is in the habit of doing this, keep in mind that there is a real danger that a lot of what you’re saying is being reworded or edited in such a way that it may end up meaning that is different from what you had in mind originally.

Usually, this problem is fixed when you set the right ground rules with a transcription company, agency, or an individual freelancer. Just let them know that you’re supposed to edit the text to make it understandable.

In other words, they would transcribe your speech as is and then they would read through the materials to look for confusing sections. At that point, they have to listen to the audio again and see which editing decision would clarify the transcription.

This is the difference between intelligent transcription and “brute force” transcription that replaces your creative authority with the editorial discretion of whoever is handling your text.

Mistake #23. Addition of Slang When It’s Obvious That It’s Not Needed

I use the words “gonna” and “wanna” a lot. This should be enough to let my transcriber know that these words should be transcribed as “gonna” and “wanna” respectively.

Because if I wanted to, I could have instructed them to automatically translate “gonna” into “going to”. This is going to trip my transcriber up because I use the word “gonna” a lot. Be on the lookout for this because this is a failure of communication.

What will happen is if you are getting your dictations transcribed and you trust it to somebody who is very literal or who writes in a very formalistic, academic type of way, there’s going to be stylistic problems in the transcripts.

It’s as if two different people wrote the piece. As a writer, it should be your product. The transcriber is just an intermediary. The process is to turn your speech into text.

But if there’s miscommunication or confusion regarding how to handle slang terms, your final text is going to look weird to say the least.

Mistake #24. Includes background noise, laughter, and non-verbal cues

This is usually not a problem with intelligent transcription services. They know well enough to keep this stuff out. But if you’re dealing with a newbie freelancer who doesn’t quite know the difference between raw verbatim transcription and other forms of transcription, this can be an issue.

Mistake #25. Repeated Sentences Are Kept in the Transcription

If there is one hallmark of intelligent transcription it is probably the ability to detect speaker self-editing. If you are hiring transcribers, this should be one of the key skills that you should test for.

When people talk, a lot of times they open their mouths and verbalize when the thought hasn’t fully crystallized yet. Things are hitting them at a rate of a thousand miles per hour and they just want to get stuff out there.

This is understandable. But as you quickly think through the ideas that you are trying to get across, you often come up with a better way of saying things. This is where the repetition comes in. I do this a lot.

When I dictate, many times the first sentence is not the ideal way of saying the idea that’s popped in my mind. You have to understand that when I write a book or a blog post, or any other kind of creative work, I only have a few lines in front of me.

It doesn’t matter if that book is supposed to be 10,000 words or the blog post is scheduled to be 4,000 words. I only have a few lines so I have to deal with ideas that come to mind based on those outlined lines.

And you best believe that the first version is not always right. Either it’s fuzzy, badly formed, or incomplete. None of those situations is good.

So I mix and match, throw it around, and kick backward and forward ideas in my head. And a lot of times the ideas come at me much faster than my ability to pick and choose the words that best match them. This is why I tend to repeat sentences.

You know you’re dealing with a very professional and highly-intelligent transcriber when they can detect your self-editing. Personally, this takes the form of me just saying what seems like the same sentence over and over again.

In reality, I’m actually editing the end. It’s the final sentence that should be transcribed because that reflects the most crystal clear version of the idea that I’m trying to get across.

Another area where self-editing is a problem involves lists. For example, I’m talking about making money online using cashback apps like Swagbucks. This program encourages you to go to different places to shop and if you buy the right promotional products you get a discount.

But as I describe the program, it turns out that there is more to it than that. So at first I would say, “Swagbucks enables you to make money online by getting cash back for every qualifying purchase you make from their partner online stores.”

Usually, I’d leave it at that but I remember that Swagbucks actually pays you to view ads, fill out surveys, that kind of thing.

This comes out as Swagbucks enables you to make money by doing a; then it dawns on me that Swagbucks enables you to make money by doing b. Then finally, it dawns on me that Swagbucks enables you to make money by doing c and d.

If you’re an English teacher reading this transcription, it looks like a mess. An intelligent transcriber would then detect the list and pack it into a tight compact form so it’s easier to read.

So the final form should be: Swagbucks enables users to make money in the following ways: getting cashback discounts, watching videos, filling out surveys, and otherwise surfing the internet.

Isn’t this version much better? Nice and tight, right?

It gets the same information across in a concise way. That’s the hallmark of a truly intelligent transcriber. The problem is when the service that you’re using just lists out the same sentence over and over with slight changes.

Not only does this tire out the reader but it makes you look like a bad writer. So there is space there for tight editing and I would think that intelligent lists are not all that difficult.

All it takes is to listen to what is being said and quickly realize that it is part of the sentence that should be condensed into one sentence instead of making it seem like a repetitive loop.

Mistake #26. Digressions and Off-Topic Content Are Included

From my personal experience, this rarely happens. Most of the time, when I dictate something I would say, “Note to transcriptionist.” When that happens they’re on notice that I’m giving them instructions.

Usually, when I dictate outlines and I come up with an insight, I would say, “Note to transcriptionist: Stop here and then create another file or put this in another section of the transcript.”

This happens because as I already mentioned in “The Benefits of Dictation Blogging”, new ideas come to mind when you are verbalizing ideas. It would be great if all these ideas are tightly connected or there is some sort of overriding theme but you can’t count on that.

That’s how weird the human mind works. This is what’s awesome about creativity. You may be dictating a review, an article, or a consumer guide on air-conditioned dog houses, and all of a sudden, you have this amazing string of ideas regarding cat food.

What are you going to do in that situation? This is a tight spot because you don’t want to let go of a potentially valuable idea. After all, it’s free content. It beats having to think about that stuff later on.

On the other hand, you don’t want to get thrown off track. What I do is I would quickly say, “Note to transcriptionist: Put this in another section or place it in another document.” Then I would blurt out the rough outlines of the idea.

It may be inspired like the outlines of Michelangelo’s David in a piece of rough marble or it can be just a brain fart. It does happen but what is important is I got the transcriptionist to get it down in writing so I can deal with it later.

At that point, I go back to where I was in the article on air-conditioned dog houses and keep nailing down the ideas based on my outline. This is not necessarily a bad thing.

But I understand that other writers might have not established a method for communicating off-topic or off-track thoughts to their transcribers.

At that point, everything is put in the transcript. Good luck trying to cut through all those different strands of thoughts, ideas, and half-baked insights with a machete.

It is not surprising that when all that stuff is left in without any kind of organization, it reduces the overall impact of the content.

Mistake #27. Failure to Follow the Basic Rules of Spelling, Capitalization, and Punctuation

Intelligent transcription should be intelligent. And the bare minimum for this, of course, is to follow the basic rules of spelling, capitalization, and punctuation. It also has to follow grammar rules. In fact, there is no excuse in this context.

People can be forgiven if they are doing a raw verbatim transcription. If you’re transcribing somebody who is borderline illiterate and doesn’t care for the basic rules of grammar, the transcript will be rough but will be perfectly understandable.

Not so when it comes to intelligent transcription. As a client, you’re paying extra for the transcriber as well as the transcription company to have some sort of quality control system.

This way any errors involving spelling, capitalization, and punctuation is removed from your text so you won’t look bad. This is not much to ask. You should filter hard for transcribers who cannot follow these basic rules.

Mistake #28. Missing Words and Passages That Affect The Overall Coherence of Transcription

You have to understand that transcripts operate at two levels: context and content. Both have to be present for the transcript to make sense and to do the job you want it to perform.

Sadly, this doesn’t happen all the time because when a transcriber is under a lot of stress and pressure, it’s very easy for them to take shortcuts. I can understand this.

But the problem is you might think that condensing certain passages or leaving out a sentence here and there isn’t going to do much for the overall flow and content of the transcript. That is too much of an assumption to make.

In fact, it’s not a decision for you to make as a transcriber. That decision can only be made by the content creator which is the speaker. If you’re reading this and you’re looking for a transcriber, understand how important this is.

This is why you have to clear this up with the transcription company, agency, or the individual freelancer you’re thinking of hiring. What policies do they have? Do they have a quality control standard that ensures that this doesn’t happen?

The problem here is that these people are not mind-readers. They don’t know that you actually place heavy importance on one passage. If they touch that passage, the overall effect of the piece (at least, as far as you’re concerned) goes down the toilet.

You have to communicate clearly with your contractor to make sure that they do not cut out anything that affects both the context and content quality of the transcript.

Mistake #29. Careless Paraphrasing

From time to time transcribers can justifiably paraphrase. As I mentioned in my list example earlier, I have the habit of spitting out complete sentences when I’m actually listing items.

I could say, “He had pet dogs, he had pet cats, he had pet rats, he had pet parrots.” That could all be boiled down into a form that cuts out as much of the phrase “he had”. It could have been, “His pets included dogs, cats, rats, and parrots.” You get the point.

Paraphrasing is needed in certain cases but when it’s excessive you end up twisting the words of the speaker. This can lead to all sorts of trouble. At the very least, it changes the context. At the very worst, the overall value of the transcript goes up in smoke.

It’s much better to limit paraphrasing to where it makes the most sense. A lot of times, paraphrasing is effective when you’re dealing with somebody with certain verbal habits.

For example, I’ve issues with lists. I also have issues with saying the word “so” at the beginning of my sentences. Depending on which side of the bed I wake up on, I also have a problem with using the word “well” when starting a sentence.

Paraphrasing can make sense in those situations but you have to be careful that the transcriber you hire for manual transcriptions does not become overly aggressive with paraphrasing. If anything they should study how you normally talk and anticipate your habits so they don’t become problems.

Mistake #30. Letting Your Emotional State Dictate How They Transcribe Your Speech

When you’re speaking out a blog post, book, article, or novel, a lot of times you are in a certain emotional state. For example, I’m not all that excited when I dictate about make- money- online types of articles.

A lot of the time, this shows up in my voice. It seems distant, bored, even distracted. But no matter how my voice sounds, I’m still thinking through the concepts and making important decisions as to which words to say. Put simply, I’m still on the ball.

With other types of content, it’s easier for me to get excited. For example, when I dictate novels and there is a battle scene, I really get into it. I talk about the weapons drawn, how hot it was; which direction the wind was blowing and all sorts of details to pump up the emotional urgency of the scene.

In both of these situations, you can physically hear the emotion in my voice but it’s a bad idea to make editorial judgments based on my emotional tones. Sure, when I’m dictating my one-thousandth article on how to make money online using cash apps, I sound dead inside.

In fact, if you were looking at me and I was dictating through Zoom, you can see the tombstones in my eyes and the ash in my tongue. That’s how dead certain topics are to me. But none of that should affect how the transcriber approaches the words I’m dictating.

The ideas are still there. I would hope that the sentences are constructed well enough. Work with that. The problem with some transcribers is that they get excited. I get this. They’re empathetic people. When they’re in front of somebody, they feel what that person’s feeling.

That’s a great skill to have for face to face interactions and customer service and sales-type situations. It doesn’t pay all that well when it comes to transcriptions though. Focus on the text of what is being said, not necessarily on how it is being said.

What a lot of transcription customers such as yourself fail to realize is that when you hire a transcriber, they are constantly making all sorts of decisions on how to process your words. This is unavoidable and a lot of it is subconscious.

This is why you must control your emotion. Try to manage the tone of your voice when you’re dictating because at a subconscious level it can impact the actual transcript. It’s important to remember that emotion is a factor.

But with that said, as far as the transcriber goes, focus on the words. Don’t focus so much on the emotion because you‘re not being hired to do raw verbatim transcriptions. Angry tones are important when you are transcribing the words of a witness who died and their testimony is crucial to a murder case.

But when you’re transcribing an article on making money with Swagbucks, those concerns are not important. Your focus should be on the text.

The Final Word on Common Manual Transcription Mistakes

As you can tell from this massive list of the 30 common mistakes made during manual transcriptions, human error is always a big risk in the transcription process. You might be even tempted to order automated transcription.

Thanks to recent developments in artificial intelligence, automated transcriptions are no longer the stuff of science fiction. If you pay close attention to the voice portion of Google Docs, it’s nothing short of scary.

Still, a lot of writers would be better off dictating their work and having it manually transcribed. Personally, I don’t think human transcribers will be out of jobs anytime soon.

While machines may be great in dictating the meaning of specific words as you string them out while you talk, they’re lousy when it comes to punctuation. They also cannot paraphrase when needed. Forget about context.

To say that automated transcription systems, whether through apps or some sort of an online interface, need work would be to put it lightly.

Human transcriptions are here to stay. If you are on the market for this type of service, read all the errors above so you can work well with the agency or the freelance transcriber you’re thinking of hiring.

Once you’ve cleared up the issues I raised above, there is a good chance that you won’t have to worry about the quality of the content that you dictate.