Leaving The Day Job

Archive for November, 2007

If you're new here, you may want to subscribe to my RSS feed. Thanks for visiting!

OK, so that’s a bit of a linkbait post title and I certainly don’t claim to have all the answers or to have made a huge fortune but I am going to detail what’s worked for me so far with Facebook ads. Standard disclaimer stuff: this is what works for me, your mileage may vary etc etc

First off, I think you need a different frame of mind dealing with Facebook ads than with your standard PPC ads. With the likes of Adwords you’re wanting to find people who are searching for a product, ready to buy it with their credit card clasped firmly in their hot little paw. Facebook users, on the other hand, aren’t looking to buy anything, they’re looking at their mate’s wedding photos or attacking someone’s vampire/zombie/whatever. Facebook ads are more about saying “hey, look at this cool product, you might like it” rather than “hey, here’s the thing you’re looking for, now buy it”.

Obviously, Christmas is fast approaching and so the products that are doing well for me are gifts. I’m taking advantage of the demographic targeting on Facebook to put the ads in front of people who I think are likely to buy them as gifts for someone else. So if you were selling kid’s toys for example, you might want to target married people over thirty. Choose a product which is unusual and has an interesting photo. Facebook ads have very little space for text and a picture says a thousand words. Use the title to get attention and gear it towards the audience you’re targeting. Use the ad body to bullet point one or two key features and include a call to action.

As I’ve previously blogged, Facebook ads seem to be hooked up to some kind of random number generator where money is concerned. Unlike Adwords you don’t get told what the minimum bid is, nor do you get any idea of how many impressions you’re likely to get for any given bid. What you do you get though is an estimate of how many people match your chosen demographic profile. If your targeting is too precise you’ll end up with a couple of dozen users in your potential audience; too broad and there’ll be millions of them. I suspect that Facebook some how takes the size of the audience into consideration along with your PPC bid when deciding which ad to show. But I’ll be damned if I can find any rhyme or reason to it!

So how do we choose a bid amount? My technique so far is a fairly unsophisticated method called trial and error. I start a new ad at 10c per click and wait to see what happens. If the ad start rolling merrily away I reduce the bid and see what happens. If it stubbornly refuses to show itself I up the bid. You can change the bid at any time and it seems that that immediately affects your click cost but the decision as to whether your ad actually runs or not seems to happen less often.

As an example, I created an ad this afternoon with a bid of 10c. After half an hour nothing had happened. Impressions stayed on zero. So I went up to 20c, 50c and then, in frustration, a dollar and still nothing happened. I put it back to 50c and left it. An hour or so later I logged back in to see that I’d had several thousand impressions and one click at 22c. I took the bid down to 10c and the ad carried on rolling. So the advice there is not to be in too much of a rush to see the outcome of your bid changes. And conversely, don’t leave a high bid unsupervised as it might suddenly start draining your balance. I’ve also noticed that if you change your daily maximum spend during the day, the change won’t kick in til the following day. So if you hit your max spend by lunchtime you don’t seem to be able to restart your ads until the morning.

This is still a small scale trial and so far my ad spend is only in double figures but at least I’m making a positive return on investment. Whether this is scaleable longer term and to larger amounts remains to be seen. We need to bear in mind that the more targeted our ads the sooner ad fatigue will kick in because we’ll keep showing the same ad to the same people. I think a possible route to success is in creating lots of mini campaigns, tightly targeted, keeping a close eye on ROI and killing them when they stop making steady profits.

I’ve been experimenting with Facebook ads for the last week or so. I’ve had some success but I’ve also been left scratching my head and wondering if I’m an unknowing participant in a very random beta trial.

The basics: you get to create an ad with a tiny headline (25 characters), a little image and a short sentence of copy. You set the maximum CPC and you pick your demographic targeting criteria and you’re off.

The pros: Ads are up and running within minutes. You can get near realtime statistics on impressions and clicks. You can change the max CPC and see an almost instant change in the rate at which your ad is shown.

The demographics targeting is really cool. I’m promoting a gift product and I’ve been able to create separate versions for men who are married, engaged or in a relationship with copy saying respectively “Buy this for your wife” , “Buy this for your fiancee” and “Buy this for your girlfriend”. These targeted ads have had higher click through ratios than non-targeted ones I ran for comparison.

The cons: sheer randomness on the stats. For starters you can view impression and clickthroughs for the last 24 hours, last 7 days or all time. Now, I don’t like this rolling 24 hour window. I want to see stats for today - as in, since midnight. It’s impossible to tell how close I’m getting to hitting my spending limit for the day when I’ve got two days stats rolled in together.

There’s also a daily stats screen which shows total impressions, clicks and costs for each day. This appears to be connected to Facebook’s random number generator. Right now its telling me I had 70-80 clicks per day at a cost of $0 each day. This morning it told me my 74 clicks had cost 91 cents. It’s also showing me a handful of impressions for today which is a different handful to the one in the stats screen in the paragraph above.

Meanwhile, there’s an account page which tells me I’ve been billed $5. It calls itself the Daily Invoices list but there aren’t any invoices to view, download or print which will doubtless cause chaos if the Revenue and Customs stop losing CDs long enough to want to audit my accounts. As yet, there’s nothing showing on my credit card statement so I don’t know if they’ve picked another random number to bill me.

Just to top it all off Facebook have taken a leaf out of Microsoft AdCenter’s book and decided to randomly reject one of my ads. Very helpfully I’m informed that it contravenes one of their terms and conditions but they can’t tell me which. There’s a few reasons why it might have been rejected depending on how pedantic or puritan they were feeling so I sent an email for clarification. The response : “Your ad was rejected by mistake. Sorry for the inconvenience”. No offer to reinstate the ad and no way to do it myself. I have tried recreating the ad and I’ve tried creating a slightly different version of the ad. It lets me do that but the ads don’t run. Meanwhile, several of my other ads, which had been running quite happily, have now stopped. Upping the CPC as high as 50c per click still won’t get them going. This another failing of the Facebook system - there’s no way of seeing what is the minimum CPC required to get your ads running.

All this is intensely annoying as the first day I ran these ads I made 150% ROI albeit on a very small budget.

If Facebook get their act together this could be a really useful ad platform but right now it’s very frustrating. For a one man business it might be possible to muddle through but for a bigger company that needs to have audited accounts the random approach to billing is just going to be a non-starter.

UPDATE: Since I wrote this post I’ve been back and refreshed my Facebook ad stats and my click through count has actually gone down. In theory, this makes some sort of sense since you are looking at the last 24 hours ie I might have had a load of clicks 24 hours ago which are now out of that 24 hour window but it just highlights what a stupid metric it is.

I’ve been experimenting with the new Facebook Ads platform. My ads have been running for a bit over 24 hours and have resulted in 20 clicks and zero sales which has cost me $1.63 (about 77p). The point of this post though isn’t about Facebook ads themselves but about my perceptions of the kind of person who would click my ad.

I’m advertising a digital photo frame which I thought of as being a geeky male gadget so I targeted males 18-35 with interests listed in photography and/or computers. Hardly anyone clicked. So out of curiosity I created a second ad, exactly the same as the first but this time targeted to females of any age who list photography as an interest. Here’s the results after one day:

Click through rate for the ad targeted at women is seven times higher. I can see four possible explanations here:

1. Women are more interested in gadgets than I thought
2. Women are more likely to click ads than men
3. Women are more inclined to “window shop”
4. My dataset is way too small to be drawing conclusions from and I should go back to doing some proper work instead of blogging about my attempts to understand women

How did I react when I saw Max the student affiliate’s earnings online? I raced into the kitchen to drink some milk just so I could race back to the PC and spurt it out my nose all over the keyboard. Here’s this 18 year old kid pulling in more off one affiliate network than I’m making at nearly twice his age in a fulltime job! (yes, I know it’s only turnover not profit before anyone says)

After I’d calmed down a bit, wiped up the spilt milk and reassured the neighbours that my loud exclamations were nothing to worry about, I read the rest of his posting. Essentially he’s asking why we’re not more open about our affiliate earnings and he signs off by asking other affiliate bloggers for our reasons for either posting or not posting our earnings. That did get me thinking. My immediate reaction is to get defensive and say that I don’t want to show people my earnings and why the hell should I? But the honest answer is that I don’t want people to find my affiliate packet a little limp and inadequate.

I’ve been working seriously on my affiliate business for about six months now on an evenings and weekends basis. I’m not somebody who has an awful lot of patience and the fact that my affiliate earnings are bumping along with no obvious sign of improving is leading me into the “could give it all up now” mindset. I’ve heard many a successful affiliate describe this point in the development of their business.

And this is where the wonder of blogging and community comes into play. Reading about other affiliates’ successes and how they got where they are now keeps me going in these gloomy times. I’m sure Max and Keiron and Shoemoney all took time to learn their craft and build up their business. They didn’t just start one morning then head down the bank that afternoon to collect the wheelbarrow load of cash they’d made. So Max, thanks for posting what you did. I’m not sure I’ll want to do the same when I’m in your position but if I do ever reach that position it’ll be thanks to people like you keeping me inspired.

Merchant product feeds are a great way to add a load of products to your affiliate website quickly. If you’re using something like Affilistore you can knock up a money-earning website in no time at all. The trouble with feeds is that they tend to be all or nothing and as we all know our affiliate sites need to be niche-focused. Getting the products you need out of a feed can be helped a great deal by the command line tools which come built in to Macs and Linux.

Geek alert! This is one way of achieving the task which makes sense to a geek like me. It’s not for everyone. If you are scared of typing commands and want a point and click approach this one isn’t for you.

First thing, you need to get to the command line. On Macs this is done by running the Terminal app which you will find in the Applications / Utilities. Different versions of Linux work in different ways but you’re looking for something called terminal or shell which will probably be in a ’system’ or ‘utilities’ menu.

Next you need to find your way to an easy to use directory. You should be able to do this by typing :

cd Desktop

if you type :

ls

at this point you should see a listing of the files you have on your desktop.

Download your product feed from the affiliate network website onto your desktop. We’ll assume that it’s called feed.csv. We need to extract the first line of the file, which has all the column names in it and place it into our output file. This we do by typing:

head -n 1 feed.csv > output.csv

Then we want to search through the feed file and extract every product which matches our search text. Say we want to find all the Bob the Builder products, we type:

grep -i "bob the builder" feed.csv >> output.csv

Notice the -i option which means we are doing a case-insensitive search and the double > means we are adding to the existing file not overwriting it.

And that’s about it. If you searched for a more generic term you might need to open the CSV file in a spreadsheet programme and edit the results a bit.

This may all seem geeky but it’s very effective and very fast. I just cut a 330 meg feed file down to just 14k in less than a minute using this technique. A 14k file I can finetune by hand; a 330 meg file is a non-starter.

If you’re new to Pay Per Click advertising there are free signup bonuses available on both Google and Yahoo.

To get a £30 voucher for Google Adwords visit this Google Adwords voucher offer page and request one. This is from an ad Google are running in the computer trade press at the moment. The voucher can only be used on accounts less than 14 days old.

To get a £60 voucher for a new Yahoo Search Marketing account send me your full name and email address via my contact page.  This is a kind of refer a friend offer so, yes, I get a bonus for referring you but since you love reading my blog you won’t mind lining my pocket a little, will you?

Both are UK only offers as far as I know.