Show HN: How I turned an old book into a (barely) profitable website | Hacker News Reader

Show HN: How I turned an old book into a (barely) profitable website

howacarworks.com

242 pointsAlexMuir104 comments

"I can only imagine that going up as Google starts to recognise this as quality content and not just content farm rubbish."

Uh, no. It will go up as a function of traffic volume, so as your site starts to rank higher organically you'll see more revenue but the amount paid isn't a function of the 'quality' of the web site, its a function of what people bid to be there and what they pay per click.

Note that if it goes up really suddenly your likely as not to have your AdSense account suspended/cancelled, which after a lengthy review process they may re-instate but Google may not (according to web reports) ever replaced the ad-revenue you lost while you were in this process.

That said, it is great you are making this information available again. If such efforts create a modest return then perhaps the archivists job is once again secure.

The OP did have an advantage that not every archivist benefits from: permission from the copyright holders.

are you sure of this ? - How does one go about acquiring such a copyright?

I think you're misinterpreting what the OP means when he says that quote.

The way I read it, the OP was not saying that his AdSense payments would increase because he got higher bids for the ads due to higher quality content, but rather that as Google started to recognize that it was quality content, the PageRank of his site would go up, display it more prominently and generating more views and uniques.

I could be mistaken, but that interpretation seems to make a fair amount of logical sense.

Yeah that was my meaning. Adsense income once you're in a certain vertical (car repairs in my case) is just a factor of traffic. Traffic comes from Google Pagerank for this sort of site. So to be clearer, I'm hoping that Google's algorithms decide this is quality content and give me a bit of a boost - at least that was my understanding of the Panda update (admittedly ages ago).

PageRank is based on links, not on content. His PageRank will only go up if he gets more inbound links to the site.

http://en.wikipedia.org/wiki/PageRank

I think the point was that Google spend a lot of effort to identify quality content vs. link farms, auto-generated content and low quality adsense-optimized content.

yes, but PageRank is a specific term with a specific meaning. It is a rank based on the number and quality of links to your site.

To say that PageRank will increase because the content is good quality is wrong, and I was just pointing that out.

What he meant to say was "... the search engine rankings of his site would go up..."

You're right, I misspoke when I said PageRank. The gist was the some though: higher quality content -> better search rankings (either through links due to higher quality content or google's non-PageRank algorithms).

Sorry I wasn't clear - I meant that traffic to the site would only go up as Google recognises meaningful content.

No worries, organic growth is a function of longevity. So to some extent you don't have much to do. Of course if you want, you can sign up for AdWords and get something like a free $100 credit just for signing up. Then you spend that by buying keywords that people might use for a query that your page would answer. That is called 'search engine marketing' (SEM) and once you burn through the free stuff you'll have to decide if its worth buying more ads or not.

Questions:

Why are there comments on the page?

How difficult would it be to compile it into an ebook and make it available on kindle/itunes/wherever else? Seems like you'd only need to make a couple of sales for it to compete with the adsense income. Maybe have it so if the visitor has read 3 or more consecutive pages it shows up with an ebook splash page, "advert free and easy to read blah blah blah"

I would buy this as an epub.

I put in the comments because I thought it'd be an easy way of adding some content for free as people add to the information already there, and also to let people moan about errors so that I can then go in and fix them. Comments are a fairly low-maintenance feedback system, especially with Disqus dealing with the spam filtering for me.

An eBook is a great idea. I just worry that the information is slightly out of date for people to actually pay for it. I know virtually nothing about ebook formats so this would be quite a learning curve for me. I'll bear it in mind though. I could realistically see 200 x $5 sales per year, which isn't to be sneezed at.

> I know virtually nothing about ebook formats so this would be quite a learning curve for me.

Pandoc can convert markdown into epub. Put it on Amazon's KDP, Barnes and Noble's PubIt, and then use Smashwords to target the other markets.

Yeah but comments will add load time and clutter to the page. Also I'll bet if this picks up any steam, the comments will devolve quickly without technically being spam.

Disqus loads in after page load, so it shouldn't add load time to the content. Also keep in mind that comments add SEO (and can be extremely valuable for this reason alone).

Plus I don't think it's such a bad idea from a reader's perspective. Seems like the worse case scenario is that no one uses it, and the best case scenario people are able to contribute to improvements.

If the comments load in after the page loads, crawlers are unlikely to see them, so you're not getting much SEO benefit.

Worst case is off-topic comments that take time to remove, significantly worse than nobody using it.

Google definitely indexes disqus comments.

They actually have a deal with google where all of their comments are directly indexed and attributed to the sites (likely similar to canonical references).

It used to be a problem a couple years ago, and sites would then use the API to load in comments and include them server-side. Obviously this wasn't ideal, but it's a solved problem now.

Just a thought: I might consider a different monetization strategy than AdSense. The ads really look awful and make the site seem significantly less 'quality' than it actually is.

I might experiment with some sort of freemium model; maybe for free you could get the first half of sections for every chapter and then you can pay ~$5 to get access to everything.

Or, as someone above mentioned, compile it into an ebook and have that advertised on the site for ~$10, but make the entire website free.

You are right, the ads look absolutely shit - I could perhaps go with Buysellads when traffic picks up.

Interesting idea - I'm a huge admirer of things that follow that model - Michael Hartl's Rails Tutorial and Avdi Grimm's Object's On Rails.

I am hoping that people fixing cars are more transactional (read likely to click an ad) than people looking for coding tutorials. I might leave it as ads for six months and then try six months of freemium access and compare the two.

I like the idea of A/B testing the ad v. freemium models, but make sure to take into account the growth in traffic that will happen in the first 6 months etc.

I'd welcome any suggestions on where to go next - it's a bit of a fire-and-forget side project, but I'd still like to get it as good as possible before I leave it to stew for 2013.

I'd load it up with all the "semantic web" style things I could find. http://schema.org type of stuff.

You could also see if it would be good as a source for citations in the relevant wikipedia articles, but that's a whole project in itself probably.

I had a look at the Schema stuff - I actually use it on another site of mine ( http://thebigeat.com ) for listings of takeaways, but it has made no difference SEO-wise and it tends to make my markup look like shit. But nevermind A/B testing, I am doing A-Z testing here - throw everything and see what sticks.

I'm hesitant to go adding my site as a citation - it's a good idea but it just seems a bit black-hat for me. Hopefully it's something that might happen organically though.

Marching in and adding links to your site to WP is certainly a bad idea, as well as being plain ethically wrong. I was thinking more of turning up to the talk page for the car article and just offering them the content, but even that probably contravenes wiki-core directive #5749327.

I mentioned it because your site seems like a genuinely good source that WP would be happy to use, making it a win-win. Don't try and use wikipedia for pagerank boys and girls.

I've always had the impression that actually being linked from wikipedia isn't necessarily a good thing SEO wise. That is I think that a long time ago google decided to penalise sites being linked from major wiki's to stop them being spammed to death.

But I'm basing that only rumour and supposition so certainly don't take it as truth.

Is there any evidence that this actually affects SEO or rankings?

Wikipedia is a perfect storm of pagerank so unless google have specifically made an exception for it, links from there must be pretty valuable. IANASEO.

For exactly this reason Wikipedia adds "nofollow" attributes to all external links on Wikipedia. You will get no SEO / page rank link juice from adding links to Wikipedia. You will probably get organic traffic though.

Duh. Can't believe I forgot about that. Not really on the ball today. What am I even doing here? This is meant to be my holiday...

Just because a link is "nofollow" does not mean you get no juice. Being linked to from wikipedia definitely helps a site's rankings...

Argument that "nofollow" = "no seo juice" is debatable. Many SEO believe that lots of links from Wikipedia does help.

I have no data of my own to support the above. Theory says that "nofollow" will not help is SEO.

All externals links on wikipedia use the no-follow tag to prevent that type of spamming (http://www.searchenginejournal.com/all-wikipedia-links-are-n...)

>I'd load it up with all the "semantic web" style things I could find. http://schema.org type of stuff.

Why though? I think the manual "semantic web markup" is a failed promise, more than anything.

The idea was flawed from the start for general web use (for one, you cannot trust content providers to add relevant semantic info, either because it's in their interest NOT to do so --e.g SEO spam--, or because they are not competent or lazy). Plus, the tools never materialised.

It's another case of "worse is better".

I don't know what your ad income is -- and I'm not asking -- but just looking at the HowIDidIt page, this is a book that screams for you to redo it for the iPad in iBooks Author. It's exactly the kind of book that app was made for.

EDIT: And btw, eBooks are my thing, so you really should listen and do it. "The Classic Back in Print for the Tablet (or iPad) Age."

I think this is looking for an iPad app too. Each section could be a low cost in-app purchase. Add a preview for each section -- perhaps an article or two from each section, for free.

I'd go on https://leanpub.com/ and turn it into an ebook.

you have starter pictures on the cooling page. otherwise, i love this - i'll show it to my kids!

I run a similar site on a different niche, which I have monetized to good effect.

1) get a few links. google will likely send you more traffic if they see others linking to you.

2) ask people to link to you on the site.

3) add some social buttons ("share this on facebook" and "+1 this"). Also, try to get it on pinterest. You might create a "5 car tips" image that could get shared as an image on pinterest, generating traffic and likely getting some blog links.

4) split test your adsense. use all three units, plus link units. then split test which ads work better when placed where. Think long term. 10% better RPM will add up. You probably want to add a skyscraper.

5) add links to amazon (through their affiliate program) where useful. need a tool for that? buy one here. I like to link to searches for products, rather than actual products. That way, the user sees all the available products rather than just the one you selected, and if that particular product goes away (which it might, because in a niche, you are probably linking to products sold through amazon by others - and they can change often - or be out of stock).

6) add content occasionally. google likes "fresh" content. changing things around, adding new articles, etc. will help you over the long term.

7) tailor your html titles to describe the content, and don't have "how a car works" in every page. It makes sense, but you really want to tell google what's there, and adding the site-wide keywords might dilute the message.

Just to add - split testing your adsense can really be a big deal. For something this small, you probably don't need to split test the colors. But, the ad placement (and which ad units to use) can make a HUGE difference. I like to select a few different formats, make adsense "channels" and then weighted-randomly select one of my splits. In other words, keep testing, but display the combination of ad units that works the best, more often. Start off at even, then revisit things 6 months from now, tweak, revisit in another 6 months.

This is the first time I've seen someone suggest split testing the ads that run on their own site. Are there any particular resources that you would recommend for this?

It's not too difficult. Suppose you have a 728x90 and 300x250 and want to see if a 160x600 would be better than the 300x250 unit. Create four new ads. 728x90 channel a1, 728x90 channel b1, 300x250 channel a2 and 160x600 channel b2. On your website half the time show a1, a2. Half b1, b2. Later go to AdSense and view stats by channel and see which made more.

Since there are so many articles, I think a search box on the left side under the introduction content would be handy.

Something that shows suggestions/auto-completions by matching the input against article titles.

Yup, search is a solid winner.

Not sure if Google T&Cs preclude you from using alternate advertising methods. If not, one approach may be to solicit ads from mechanics, car shops, car maintenance training schools, etc. and build in some geo-awareness so you serve up ads local to the site visitor. Could be a very high conversion/CTR and be able to charge a nice fee.

You missed a sheet reference on this page: http://www.howacarworks.com/articles/lubricating-hinges-peda...

Aha thanks, a sneaky change from See SHEET to consult SHEET.

Assuming its a Rails or Sinatra app, building an iPad app based on JSON output would be fairly easy. I could easily see you making some extra revenue by selling it for $0.99-2.99

This can be a really useful website that could potentially earn you _interesting_ money.

I'd take your first few $15 and reinvest it by getting someone on oDesk to go through and fix stuff.

* First section I looked at was 'Brakes'. First article is about the windscreen washer. Then there's two articles called "Adjusting the Brakes /1" and "... /2". Why two?

* In the first there's one massive unreadable paragraph. In the second there's paragraph-length titles.

* I'd also look at getting your oDeskers to put the images in interesting places within the text.

At the moment it feels like an OCR dump of a book rather than the amazingly useful website it could be (and thus it feels like an SEO AdSense site).

Thanks that's a great idea. It's a little more work than putting the job up on oDesk though, as I'd need to implement some mechanism for people to edit and update, and for me to vet it.

The truth is, having scanned and OCRed it all, I'm sick of the sight of it for the time being so I haven't proofread it. I will get onto this. And your suggestion about moving the images around is a great one - I'm conscious that these pictures are lovely but I'm not displaying them well at all.

You could simply pay a bored technically-minded university student such as myself to do the work :)

Put some contact details in your profile and I'll be able to contact you.

Done!

Paragraphs would be nice. And proofreading.

For example: http://www.howacarworks.com/articles/removing-the-oil-pump

"On en gines where" " If yot are in "

No wonder Google doesn't pick it up.

I would

1) Make the pages editable by trusted individuals but with your validation before any change goes live. Allow the authors to be credited (with links to their blogs, pages etc etc).

2) Why stop at cars? Find a domain "howdoesa???work" and subdir it off that - offer written credit to anybody who wants to add any large-ish content.

3) Put in some clever links to amazon etc so that if someone wants to buy a new car battery you might be able to nail som e affliate fees. Especially if they are on your page, have identified that they need a battery - why go back to the search bar when you have "click here for battery offers". I am suggesting keeping it "neat" though. Please don't start adding horrid looking ads disguised as CTA buttons.

4) Cut your mate into a %age so he's a good posterboy for digitising other content.

> Make the pages editable by trusted individuals

Way too much work

> Why stop at cars?

You'd need some content to begin with. He was lucky to have a friend who gave him permission to use the book. Other material might not be so easy to get.

> Put in some clever links to amazon etc

Great idea. Done tastefully, it would not be intrusive like an ad, but a useful tool for the reader who would need this kind of book.

I'm surprised no-one has mentioned how badly in need of proofing this content is. I just looked through a handful of sections but the ones I saw were littered with missing words, out-of-place headings, and poorly-cropped illustrations. It seems like great content, and the idea of charging for it has merit, but you're not going to be able to do so without proofing the content; it doesn't seem to me that even a cursory scan-through of all the pages has been done.

I've always wanted to see the entire library of haynes manuals put online or even in an ipad/iphone app. This looks a bit like a proof of concept to it. Perhaps the OP could approach Haynes pitching the idea and showing this as an example?

>approach Haynes pitching the idea and showing this as an example?

This is the correct answer. At least try and get a friend to do it, and take a cut.

The idea of being able to flick through all the haynes manuals on an ipad is brilliant. And I don't even like cars.

Question to motorheads:

Would you use a shop manual exclusively on an (your) ipad?

Or would you prefer to have a paper copy, so that you don't get your ipad greasy and scratched while you're working on a car?

Question to everyone: when are we going to have devices that are cheap and rugged enough that my first question doesn't matter?

Edit: If that ever happens, they won't be sold by Apple, the company that made unboxing a fetish.

No. I need a paper manual. I can write notes on it with grease, and not care about. It doesn't get scratched, need updates, or charging. Plus I can throw it in the trunk of my car and not worry about it.

I tend to get PDF workshop manuals, print off the pages I need for a task and put them in plastic pockets to stop them getting destroyed in the rain.

see Alldata or Mitchell on Demand. Both are web based auto repair manuals. I sold Alldata in the 93-95 years and Mitchell OnDemand from 96-2002. Most auto repair shops have them. Typical use case is to print out the relevant parts, take the paper to the car and use it make notes, trace wiring, etc, then include the notes with the repair order.

Haynes has manuals online already. http://www.haynes.com/

http://www.youtube.com/watch?feature=player_embedded&v=n...

It's not exactly flipboard, but the info is certainly there. The pricing seems a bit obscene, but if the alternative is taking it to a mechanic then I guess it's a cost savings.

This may have been a lucky break. Google keyword tool suggests that "how a car works" is 1) low competition 2) 135,000 global 3) 60,500 local. This is an amazing and rare find, most keywords like this are already well picked over by niche site creators - also from the looks of it you were the first registrant in 2011.

Yeah, I felt lucky getting the domain. But then I also loved FixingACar.com as well.

This is nifty, and I feel like this could be pushed a lot further. Epub output is an obvious next step, and I'd imagine a better table of contents view would make the web version more inviting. As for generating money, maybe hotlink products to an Amazon/etc affiliate link so when a user reads the page about "Replacing engine mountings", you can suggest sites that sell engine mountings.

I'm curious about the intermediate "Horrible HTML" stage. Did it retain the book layout at the stage? The scanned book images look much more appealing than the blog-style version (in my opinion). Even though the HTML was horrible, did it still display properly in a modern browser? I wonder if you could write a script to clean up and modernize the HTML while still retaining the nice layout, rather than reformatting into a blog and dumping images to the sidebar.

It would be interesting to see how much the process can be streamlined so you can convert books en masse. I assume there are plenty of books of this style with really good content that would be useful to convert in this manner.

Sadly the HTML rendering was fairly dreadful but it did at least have the images in the correct order. It retained no placement, and each page is very different in layout. My early thought was to have perhaps 4 layouts which I could manually assign each page to. (eg. large central image, 2-large images, etc.)

The obvious next step to me would be for you to repeat this process with another out of print book now that you've established a market and proven its viability.

Offer the copyright holder some share of the profit and scale like mad.

Great stuff!

That was my initial thought but I have to see how the revenue scales. At the moment it's only profitable because my labour has had zero cost.

It might scale quite nicely though - I think finding the right material is probably key.

I'd make more of a feature of the awesome cut away drawings - I always loved them as a kid in Usborne books. Did your friend do the drawings?

I'm not sure on the illustrations - I'd have to ask. He wrote some of the content, but also ran the company that owned the copyright on the work. They are wonderful and I have to emphasise them some more.

It's a lovely idea, but from a design point of view, the site looks really quite unpleasant: the green pastel background, ugly font choices, an uninviting table of links as the home page, and inlining the figures far too small to be useful.

I would recommend at least following some of the design choices of the original book, so the difference between text and figures is not so jarring. But better to extract drawings out of the scans and relabel using a consistent set of fonts, while creating a more interesting and captivating layout. It would be awesome to see this book brought to life by a more modern design with all the benefits of hyperlinking (both internal and external, e.g. wikipedia).

It's a real shame the presentation is so poor, as the information is actually really useful and interesting. A great start anyway.

this is great, actual useful content with great illustrations. It's refreshing to see something with some practical use.

Growing up I remember being obsessed with this: DK The Way Things Work. it was an amazing mac CD ROM with illustrations of everyday things broken down, with moving parts and all...

Probably one of the reason i became an engineer.

http://en.wikipedia.org/wiki/The_Way_Things_Work

http://www.amazon.com/Things-Work-CD-ROM-Version-1-0-mac/dp/...

Horrible Youtube video ( best i could find... ) https://www.youtube.com/watch?v=f4MIVU5i_xE

Cannot recommend it enough for kids. They don't make em like they used to :D

The Way Things Work (the book) is amazing; the new version also includes chapters on computers and electronics. Not sure about that dated "multimedia experience," but the book itself definitely holds up fantastically well.

Thanks for showing us. While my father can fix / build nearly anything, I overspecialized in computers and electronics, so I know almost nothing about cars. I've been looking for a way to remedy this and your site looks perfect.

Buy a wreck and take it to pieces and try to understand it. Then just sell it for scrap and recoup most of your money. You could even sell the parts individually and make some profit. Then try a project car - I'm surprised more hackers aren't into cars really. Sadly I'm in an apartment at the moment without the space.

Looks like your site is already number 3 on my google search for "how a car works".

The homepage is a bit boring. You have a lot of pretty pictures, you should some how use them on your homepage. Perhaps include a relevant picture with every chapter heading.

And for good measure, include all the social share buttons junk. Google+, FB, Twitter, etc... Especially for if you get popular on HN. You'll want those sharing karmas...

I would also drop another adsense text box into the side nav. Perhaps remove the one at the bottom of the page.

> And for good measure, include all the social share buttons junk. Google+, FB, Twitter, etc... Especially for if you get popular on HN. You'll want those sharing karmas

Ugh, no, for crap's sake, don't do this. People will share things if they want to.

http://www.poynter.org/latest-news/media-lab/social-media/14...

http://www.business2community.com/social-media/a-hot-button-...

http://www.searchenginejournal.com/too-many-social-media-sha...

etc.

You're right about the pictures for sure, I need to add a few random pics and thumbnails for chapter headings. I focussed on SEO for starters and Googlebot doesn't see pictures, but now that I have people I need to think about actual aesthetics.

Questions:

1) How much time did it take after the scanning to get the site and content online? 2) Do you think you could automate the process and get a 2nd book done in a faster way?

After the scanning I took about 2 days to OCR everything. That was fairly mindnumbing.

Then it was a case of coding up a fairly simple Rails app and parsing that content - another two days. Some thinking time in the middle - I'd say a week. Spread over a year(!)

And no mention of using Twitter Bootstrap?

I like how the site looks/works, you did an awesome job, but some credit in the "how I ..." should go to the Bootstrap team.

Because they need it...

You're right - and also credit should go to the excellent http://bootswatch.com/ which is where I got the themed Bootstrap from. I'm a lover of Bootstrap but I hope it doesn't define the site.

Those images on the right side of each page are great; They could be used to good effect to make the site more appealing visually as well as unique, if you displayed one or more of them in full size interleaved with the text: maybe one on top of each page; and one on the sidebar. They may also help with SEO (use alt-image tags), and ad Click-thru-rates, if placed close to or next to the ads.

The biggest beef I have is that the site doesn't tell you how a car works. I clicked on "Transmission" hoping to see how an automatic transmission works, but the TOC only had information on checking, filling, and adjusting various things. Perhaps the fixingacar.com domain would be more appropriate for this content?

Scope creep suggestion: It may be cheesy but I believe a rotating little GIF animation on the home page/header of how a small part of a car works would draw in people. As it is now, it is static.

Think more http://www.brainpop.com, less Encyclopedia Brittanica

I, for one, would be interested in seeing that nokogiri script that turns ugly HTML into markdown... although I'd totally understand if you want to keep it close as a competitive advantage. ;-) I regularly have to deal with the ugly HTML that various apps' "Export" features disgorge.

It's so hacky that it's almost certainly a competitive disadvantage.

Converting to markdown is a bit of an overstatement - it only handles the limited content that I encountered in this book.

- H1 heading -> ##

- Strip P tags and add newlines

- Replace 'See SHEET \d+' with [[Article:\d]]

You inspired me! This weekend I did my own book to website conversion: http://recipesforseafood.com/

I'll keep you posted as to what $$$ it generates (if any)...

Cheers,

Marcus

This is really fascinating. Great article. I think fiddling with paper->web (or something else digital) and improving that workflow sounds like fun.

what did/do you count as your expenses? surely you do not pay 60$/month for hosting. did you add the work you had to do?

Hosting is on a $20/month Linode along with a bunch of other projects/sites/bits of code fluff. The actual bandwidth used by this site will be virtually nothing I imagine. I haven't valued my time at all. In that respect it is already earning me around $40 a month - a meal out.

Are you paying any money to the original author?

He's not bothered - it's an experiment for both of us. If it works out then I'll see him right, but there's no point in hashing out some lengthy agreement if it doesn't bring in the bucks.

That is really nice. Congrats!

this is awesome - I'm curious though - aren't you theoretically violating the copyright of the actual author?

This is essentially a derivative work. And the blog post specifically states he had the permission of the holder who was a family friend.

Now having the permission not in writing even (especially?) of friends and family is generally not a good idea, but not exactly the end of the world either.

Cool experiment. I thought about doing something like this myself. Maybe a remix of classic books under the public domain to create a new work with a contemporary message. Maybe then publish it online and monetize it via Adsense (although I'm not a big fan of that business model).

Instead, I decided to share a personal finance hack I developed called the No Budget Budget: https://leanpub.com/nobudgetbudget/ and sell it as an affordable ebook.

Again, the monetization isn't great, but it means more people can read it and hopefully benefit from it.

The point is ideas are everywhere. I like pursuing ideas that will have the greatest positive impact. Then, like the legendary NFL coach Bill Walsh used to say, "the score takes care of itself".

Your link doesn't work for me.

Fixed. Thanks for the heads up.

Reply on news.ycombinator.com