The real costs of a ‘lite’ version

July 20, 2008


The PTB (powers-that-be) are appalled at the cost of translation of our Help systems (about £20,000 for our biggest Help project into one language). For stuff we’ve had previously translated, or stuff we’ve got some translation memories for, we can probably get that cost down by 30-50%.

It has been suggested that I could maybe create a cut-down version of the Help for translation purposes. I suppose I could exclude the Help for certain dialogs or areas of the system, but who would decide which areas were important? I could exclude the Help for certain ‘self-explanatory’ fields on dialogs, but if they’re self-explanatory, I haven’t written much for them anyway. I could exclude any procedural topics, but I’ve been adding more, in response to reports that the Help doesn’t tell you what to do.

I’m stumped, really. Any ideas? Has anyone ever had to do this?

My response

Let’s look at their bottom line figures again… You say you/they can get the cost down about 30-50% by cutting the Help file words down. So from 20,000 jellybeans (the currency doesn’t matter) it now costs them 10-15,000 jellybeans. That means they’ve ‘saved’ 5-10,000 jellybeans. But have they really?

First, there’s the cost of you finding out from others what they want cut, or making that decision yourself. Let’s say it takes a week to analyse the Help sufficiently to make that decision (you still haven’t cut anything, just analysed where the cuts can be made). Let’s say it costs the company 50 jellybeans per hour for you to do that analysis—that’s 50×40 which is 2000 jellybeans.

I haven’t factored in the cost of getting others to decide. Let’s say those others all cost 70 jellybeans an hour to employ (they are marketing types and developers after all), and they spend 20 hours each trying to decide (4 people x 70 [jb] x 20 [hrs] = 5600 jellybeans).

Somewhere in there you have a review meeting where everyone disagrees with everyone else, so you all go away and think about it and analyse it again and come back for another meeting. Let’s say two hour-long meetings attended by the 4 others and you – that’s (2 [hrs] x 4 [people] x 70 [jb]) + (2 [hrs] x 1 [you!] x 50 [jb]) = 660 jellybeans.

So assuming you can all agree what stays/goes you’ve already cost the company 2000 + 5600 + 660 jellybeans… 8260 jellybeans – and STILL you haven’t actually cut anything.

OK, keep with me here folks… Part of the cutting decision is to keep the English version in its entirety, and make a ‘lite’ version for everyone else (I won’t go into ‘them’ being treated as second-class citizens—that’s an argument for another day). Now you have to do the actual cutting. This means copy/pasting into a new document (I’m assuming only ONE help file here; there may be many more), reformatting the new document, relinking the new document, building the help files, checking the new document for broken links, fixing, tweaking, etc., etc., etc., BEFORE it even goes out for translation. Let’s say this takes you 2 weeks — 2 [wks] x 40 [hrs/week] x 50 [jb] = 4000 jellybeans to do the cut.

But wait, there’s more!

Now you have TWO documents to maintain whenever there is a change—and it’s not that easy. You have to check if the change is required in BOTH docs, or just one, then make the change. And you have to remember to do this EVERY time. Somewhere I read that to maintain two similar docs like this takes about 50% more time than just maintaining a single source. So let’s say you stay the way you are as far as work practises go. The changes to a manual that used to take two weeks now take three weeks… So instead of it costing the company 2 [wks] x 40 [hrs] x 50 [jb] (4000 jellybeans) to make changes for a new release, it now takes an extra 2000 jellybeans, and puts you further behind on deadlines to the point you might need an extra tech writer. <the silver lining for those who are currently unemployed!>

So, what did we have earlier? 8260 jellybeans to get to the cutting stage + 4000 jellybeans to do the cut + and extra 2000 jellybeans EACH time a new release is to go out the door. That’s 14,260 jellybeans according to my calculator! So it’s COST you more than 14000 jellybeans to SAVE 10-15000 jellybeans—for ONE help file!

Of course, the numbers I’m using may be spurious, and you’d have your own estimates of how long something took—I’m just plucking figures out of the air to prove the point. And the point is that unless the company REALLY sits down and looks at things beyond the price quoted for the translation, they are going to be forever stuck in this loop of the quoted price being exorbitant. What is exorbitant is the cost of doing what they are suggesting. Has anyone actually costed YOUR time to do this?

You could go with a single-sourcing solution where you could reuse content without copy/pasting etc. to create a ‘lite’ version. But then they’ll complain at the cost of the tool and the cost of converting (and baulk at the cost of training, without which you’ll spend even longer getting up to speed), without thinking of the overall savings they could be making, especially as a good single sourcing tool can also reduce translation costs because you only send the revised topics to the translator, not the entire manual. Yes, like anything new there’s always a conversion/learning curve cost, but hopefully the ROI comes about fairly quickly.

Then there’s the whole moral issue of treating people who speak languages other than English as though they are somehow lesser beings and not as ‘deserving’ of the Help…

I hope there’s something in there that you can use to support your case. Good luck!


  1. There’s a whole other set of costs that could be involved here – the cost to the company of not supplying adequate help information. That could be more support calls, and fewer repeat orders, and so on. Of course, these are items that can’t be measured today, while cutting translation costs can be measured today.

    I’m sure you know how companies can theoretically save money on documentation through planning and re-use, but it doesn’t sound like the company in this case would be interested.


  2. Excellent points, David. And, as you say, very hard to quantify.

  3. Is the company going on the full word count of the help?
    Has the source material been put through Trados or a similar translation memory system to assess the levels of reuse (from words/phrases through to full sentences)? If there is even a base level of reoccurring text, such as referring to the “whoosiwhatsit screen” or even phrasese like that, AFAIK Trados will pick that up and different translation rates will apply for existing translations. If the original 20,000 jelly bean quote is purely on word count, the source material should be analysed under an RFP by the vendors.

    I won’t even talk to the “lite” version of the help …

  4. A similar problem arose at my place of work, but the solution was slightly different.

    Our company’s help file had been created before translation was even an issue, so it was wordy, colloquial and, over the years, had grown big and chaotic. It was full of complicated sentence structures that were hard to translate, hence it cost a fortune to translate.

    Someone suggested a “lite” help file would solve translation problems but the real problem was that the original help file needed editing and rewriting anyway – especially to take the company’s new international reach into account. When I rewrote it I reduced the 50,000 help file to 36,000 words, without losing any information. The company got a better, refined help file which was cheaper to translate.

    So the solution wasn’t a lite help file, but a new help file…!

  5. A solid edit of any documentation that’s been around a while is always worthwhile. However, if you’re the author AND editor, it’s hard to do while it’s all still fresh in your mind and while you’re still close to it.

    Sometimes you have to step back from it for a few months or years before you can concisely edit it. Unfortunately, we often don’t get the luxury of time to do this… until translation forces us to do it. At that point the powers that be realise that cutting words (editing!) is A Good Thing.

    Cutting 14,000 words from your help file without losing information was a great achievement!

  6. […] I’ve written before about the hidden costs of migrating your existing content; about being careful of software demonstrations given by a vendor; about making a business case for getting what you need; and about the real costs of a ‘lite’ version. […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: