Welcome to WebmasterWorld Guest from 184.108.40.206
Forum Moderators: not2easy
I run an education-based website and have been offering users free downloads of one of our ebooks, as a promo thing for one week only. I was happy with the uptake at first. I noticed that on average around 1500 downloads were made each day. However, when I double-checked the logs I noticed that up to 80% of the downloads were unsuccessful. Visitors had started downloading but either given up or been unable to download the files. How can I make the experience more successful for visitors? These were only small-ish files, between 5-6 MB (.PDF files).
First thing you might want to do is show the user how large the file is by putting the file size next to the link. Another option may be to offer the PDF as an alternative option, perhaps a compressed (.zip) file download. The compressed version should significantly reduce the PDF size and be a welcome option to those in-the-know.
When my sister was living in a rural area, she couldn't download files that big, or even close to it. Even if she tried starting the download late in the evening, and leaving the computer running, she'd find the connection had been dropped some time in the night. It was simpler for me to download whatever files she needed, burn them onto a CD, and mail them to her.
5 megs might not be that far of depending on size and how graphically intense the book is. We made a pdf version of a magazine that we put out for the site and a 40 page graphically intense article came out to 4+ megs with images saved at 96 dpi.
I tried zipping one of the files (about 3.6 MB) with WinZip but it came out at a similar size (about 3.5 MB) which doesn't really help. Am I missing something with Winzip? In my experience it's great for zipping up multiple files and sending them or posting them online but I haven't found a way for it to significantly shrink the files. Any ideas?
That lets you control how big you want the file to be, versus image quality.
So you could provide your users with a choice of high quality (big file) or low quality (small file).
(As Acrobat already compresses the images, it's not surprising that WinZip doesn't succeed in compressing them any further.)
The problem with new versions of acrobat reader and tyeh IE plug in is that in IE the user has to wait for reader to load, THEN reader has to read almost all the file before the page comes up on the screen. This is crazy. It was so much faster in earlier versions, where page one appeared firsrt and THEN the other pages started loading in the back ground.
Also - have a choice of download servers.
instead of asking "... them to "right click" and select "SAVE AS" ..." you could use those force-download tools that prevent documents to be opened directly but show the "Save as" dialogue. With the download window open and the progress bar showing the users are imformed about the status of the download.
It looks like I will need to update my Acrobat to get the built in compression (I've got Acrobat 5).
I'm starting to realise that the issue is with the pdf opening up in the browser, rather than users being able to just download it. I guess if I zipped it so it was "worksheets.zip" rather than "worksheets.pdf" that problem would be solved and more downloads would be successful because it would simply start downloading... Is that the case? The problem with that, though, is the inevitable batch of "I can't open the ZIP file..." emails coming in, which isn't ideal!
Or ask users to "right-click, save", but most people when they see a link, click it.
The files are already chunks of a larger file, so I don't want to split them much further, otherwise users will be downloading the book page by page - I guess that is an option too, but time-consuming for the users.
Many thanks to all again for the inspiration!
Then I designed an intro-page in html, where I put that hint to right-click-and-save the file and also a table with estimated download-time for various connection speeds.
Finally I used pdftohtml to provide an alternative html-version with minor quality for those who are interested in a first quick look-at, and placed a linkt to that html-index-page on the intro-page. If I remember correctly, pdftohtml did not work with the original version, because of the cmyk-format of the images or so, but it worked after shrinking and saving it with gs-view.
Let me add that googlebot initially had problems to swallow that pdf, but I recceived no error-message in my sitemaps account for months now. 70MB of unique content!
Our main problem now: We receive a lot of orders via fax from people who only tell us the product-ids from that catalogue, though we actually don't store those products;)
So, it seems an amazing number of people is using that download. The german telecom is currently working on a 50Mbit-highspeed infrastructure. I can see the days where my 10/100MBit Ethernet-Hub ist THE bottelneck.
It makes me think the best way would be to save the original Word pages that I used to create the PDF as html pages and create an html menu so that people can easily navigate it like a website, rather than having to download anything - and print the pages they want.
But it's frustrating when you think how easy it is to use PDF files and how good they make the pages look, with their correct formatting - how functional Acrobat Reader is, and so on... Yet there are still so many aborted downloads. Very frustrating! But I guess we always have to consider our users' needs rather than think "Well I would be fine with it, so why can't they get the hang of it...!"
Back to the drawing board I think!
I use the photoshop function to save for the web quite frequently which save through ImageReady or something. This could cut down on image sizes a lot. This was already suggested in a round about way.
Also, on all our servers we use mod-gzip which cuts down on bandwidth and I think it cuts down on load times for large pages.
The server zips the file - transfers to browser - then browser decompresses....only problem - I don't have a clue if this can work with PDF file types.
[edited by: Bilbo123 at 2:00 am (utc) on Sep. 22, 2006]
Yes. Generally you assume one single page to cover 4k mere text. That means your documents would conist of more than a thousand pages each if they were 5 MB, which I doubt. This clearly shows the nonsense to use the pdf-format for your own documents on the web. Besides, zipping pdf-files normally has almost zero effect.
You may use some css-formatting for page/text width and form-feeds if you try to design a printer friendly version of your html-documents. I used that technique for my invoices and it works perfectly fine. Your only problem would then be to tell your visitors they'd beter switch off automated page-numbering in internet-explorer. Took me five years to find out this is possible;)
In Acrobat Reader preferences (writing on OS X here), choose 'internet' from the sidebar. there is a box labeled "display pdf in browser using...". Click on it to uncheck it, to prevent pdf's loading in browser.
Wouldn't that change the setting for your own machine, rather than the end-user's machine?
The better solution here, seems to me, is to force the file to download with a 'save this file' dialog (using a mime-type setting? I can't remember).
zipping pdf-files normally has almost zero effect
Providing PDFs obtained from multiple sources, I noticed during rsync (home to server) that some of the PDFs had astonishing compression. At that point I tried WinZip on a test bunch of PDFs and noticed that some shrunk by more than 90% (more than 10-fold reduction). My policy now is to zip every PDF. Admittedly, many have less than 5% reduction when zipped, but that is still a reduction in both bandwidth and download speed - remember, gzip compression cannot be used on PDFs since it shafts most Windows display-in-browser.
A note on forcing download-menus:
Both 'Content-Type' and 'Content-Disposition' must be set within the headers, in that order.
for MSIE & Opera: "Content-Type: octetstream"
for all others: "Content-Type: octet-stream"
else: "Content-Type: application/force-download"
then: "Content-Disposition: attachment; filename=filename"
(no quotes throughout, including the filename; filename must not contain `Web', `Browser' or `WebBrowser'!)
(conversely, if you want it to be inline, use the correct mime-type for 'Content-Type', and change 'attachment' to 'inline')
the original Word pages that I used to create the PDF
Okay, we have part of the equation here at least. Now we know the original is in Word - what software exactly are you using to create the PDFs? Acrobat Distiller? PDF Writer? Another app?
I ask because a friend sent me a large Word document a while back and when I asked why he did not send as a PDF he said it was because the third-party PDF utility he was using did not produce any appreciable decrease in file size.
I took the Word doc and PDF-ed it myself (with Acrobat Distiller 4) and cut the file size significantly (originally around 2Mb Word doc, cut down to about 400kb as a PDF).
The same document, had it been created in Adobe InDesign probably would have been around 200Kb when exported to PDF. My point is that some apps create more efficient PDFs than others.
- HTML (i.e click onto a web page) (averaging about 20 KB)
- Word .doc (about 30 KB)
- .pdf (about 90 KB)
- .txt document (about 1 KB)
This way everyone should be able to get hold of the material that they want!