Online Advertising Answers: September 2020


Hello, Justin Driskill bringing you the best of my mailbag this month, with Online Advertising Answers. To get your question answered (and possibly featured here) find me on Quora, and ask away.

This month we delve into some HUGE questions… if you work in digital marketing. Where did it all start? Am I succeeding on Facebook? How do I encourage clicks on ads on my site? You know, normal stuff.

 

Online Advertising Answers

 

This month’s questions:

[Click these links to skip down to the question]

 

Let’s dig in…

 

 


 

Q:When did online advertising start?

&

A: “Hi Mohamed,

Online advertising arguably started with a spam email, the first of which was sent on May 3rd 1978 by Gary Thuerk of the Digital Equipment Corporation[1]. It was an invitation to a product demonstration, and it was sent to 393 people.

It was in all caps because that’s all there was back then…

 

The first spam email

 

While it wasn’t personalised in any way, it was segmented – it was only sent to West Cost technology enthusiasts.

Back then 393 recipients was an incredibly long email list. Although this email was sent a whole 6 months after the BCC field was invented[2] it was not used for this email. This meant that the actual email sent looked more like this:

 

First Spam Email with addresses

 

Needless to say, it was not well received. The online rage that followed this unsolicited mail meant that there were no further (recorded) attempts at spam for almost ten years after this[3]!

[If you want to read some of the reactions, they have been catalogued here. They include some discussion of freedom, forwarding job offers, and dating sites. Remember – it was the 70s, and these were academics]

Just imagine doing something so annoying that no-one else did it for ten years!

 

Online Display Ads

According to internethistorypodcast.com, the “first true, linked, clickable, paid advertisement on the world wide web goes to a website called the Global Network Navigator, or GNN, which, in 1993, sold a clickable ad to a Silicon Valley law firm, Heller, Ehrman, White and McAuliffe.”

However, as there seem to be no interesting details about that first deal to be found, let’s ignore it in favour of the popular myth of the first online ad, which is much more interesting and much more detailed….

 

Have you ever clicked here?

According to most of the internet, the first clickable online ad ran on on October 27th 1994 and looked like this[4]:

 

The First Display Ad

 

It was run by AT&T with the launch of hotwired.com – which was the precursor to wired.com. There were apparently a set of 14 ads from different advertisers which were all set to go live, AT&T just happened to go first.

The AT&T ads were part of a larger ad campaign called “You Will”[5] which featured lots of futuristic ideas (of the time) such as GPS and eBooks.

This is why people latch onto the AT&T ad as the first banner ad even though it’s one of a set, as it seems to have predicted the rise of internet advertising and so makes for a good story.

 

What size was the first display ad?

It was meant to be 460×60 – which is only slightly different from the first standard ad size of 468×60[6]. However, browsers at that time automatically put a two-pixel blue box around images (!) so the actual size was 546×56.

These ads were non-animated gifs with limited colour palettes (usually of 2 or 3 colours). This was so that the ads had the smallest file sizes possible – as dial-up modems would have taken a while even to download a simple coloured box.

Also, most computers at the time didn’t have graphics cards that could take lots of colours!

 

Where did the display ads go

If you clicked on the AT&T ad, then you were taken to this landing page[7]:

 

The First Landing Page

 

As this was a brand new idea, most of the advertisers didn’t have websites themselves yet. They were just so excited by the idea of online advertising that they jumped in without thinking it through – there are countless stories of late-night phone calls from brands to “technical geniuses” to get a website built asap!

To get ahead of technical issues, hotwired insisted that all the ads go to landing pages on the hotwired site itself meaning that they weren’t just buying ads, they were also getting three pages of webspace as part of the deal.

How much did these display ads cost?

The banner ads (and three webpages) cost a flat fee of $30,000 dollars for 12 weeks. There were no promises made about performance, as no-one had any idea how they would do.

The ads were sold as sponsorships to different sections of the site (people would see the same ad over and over whenever a page in that section loaded). This must have seemed like a great deal as in 1995, hotwired.com put up the price, and other websites who started offering online advertising charged about the same.

If you ever wonder where the dotcom bubble started, then this is a good place to look.

How did these display ads perform?

The ads apparently started off at around 70–80% CTR[8], then the AT&T ad (which was apparently the top performer) averaged out about ad a 44% CTR over time.

This level of CTR is pretty amazing compared to modern display ads. Today, display ads on average get a CTR of around 0.2%[9] (which is over 200 times worse):

 

CTR Benchmarks

 

However, there were literally no other ads running at the time. No-one had ad-blockers, no-one was suspicious of online ads, and there as no ad blindness. Quite the contrary – people were probably excited to see a banner ad, which is not how you would describe modern reactions to online ads.

 

"Hey Look, a Banner Ad" Simpsons Quote

 

Craig Kanarick, who designed the AT&T ad said:

“I would like to think that it was because I designed something that was incredibly beautiful, and that our copywriting was so amazingly compelling, but I think the truth is that it was also a time when people were not sort of inured to ads. And not angry about them.”

On top of all this, I would be shocked if the tracking of this ad was anything except for terrible – as no-one had ever tried tracking something like this before.

And you also have to remember that not many people were even online in 1994. There were an estimated 2 million computers connected to the internet[10] back then (compared to 4 billion today), so a high CTR for a small audience is not that surprising, nor will it have brought in that many actual clicks.

 

The legacy

As well as the deluge of online ads we see today, and the dot com bubble which occurred when people started massively overvaluing their own websites, the first ads also kinda drove the invention of modern servers…

Jonathon Nelson, who worked on the first campaigns that ran on Hotwired said:

Most companies, that had a little bit more money than we did, they would essentially run one web server per site. And so, if you had Volvo, and Club Med and AT&T and MCI, which we were all hosting, at the same time… We didn’t have enough money to have six different servers for six different sites. So, we hacked the original CERN source code so that we could host multiple websites on a single box. On a single PC. We became sort of famous for that hack. And people were asking us for copies of it. And then we said fine, we’ll put our patch to the CERN web server. Hence, the joke, A “Patchy” web server… we’ll put that up in the public domain. (…) And that’s how the Apache server developed. And Brian Behlendorf, and Cliff Skolnick, who were two of my partners in Organic, really went off to develop the Apache server. And now it’s got, what, 60-70% of the Internet server market?”

So there’s that.

 

Read on

I never say this… but read the footnotes! There are some weirdly great histories of the first ads out there and its entertaining to see how excited people were to jump on board this new medium. It’s also fun to read how optimistic people were about the future of the internet, and how technically challenging putting together a single unformatted web page was back then.

I hope this helps,

Justin”

Footnotes

[1] Origin of the term “spam” to mean net abuse
[2] Who invented the concept of Bcc in email?
[3] The History of Online Advertising
[4] The ‘First’ Banner Ad
[5] You Will – Wikipedia
[6] 468×60 Definition (Banner) | The Online Advertising Guide
[7] The First-Ever Banner Ad on the Web
[8] On The 20th Anniversary, An Oral History of the Web’s First Banner Ads
[9] CTR Definition (Click-Through Rate) | The Online Advertising Guide
[10] 1994 in technology: What the Internet, computers and phones were like 20 years ago

 

[This originally appeared on Quora]

 

 

 

Q:What is considered a good click-through rate on Facebook ads?

&

A: “Hi Roksana,

The average CTR of ads on Facebook is about 1% so generally speaking, you should be aiming for higher than that.

 

CTR Benchmarks

 

However, your goal should always be to beat your past performance. The best benchmark you can ever have for your own ads is how well they have done in the past. This is because the combination of your brand, product, and the person running your ads is unique, and so your performance will be too.

Taking your CTR from the previous month and trying to improve it is the only real path to success.

 

[Need help? You can calculate your CTR by using this free calculator]

 

How Do You Improve Your CTR?

Optimising your campaigns is arguably the most important part of running ads. For Facebook Ads you should look at the various reports to see what is working and what isn’t.

Here are a few ideas on what you can optimise towards:

Please note: For each of the below only look at things which have delivered over 100 impressions. For example, if an ad hasn’t delivered 100 impressions over the past month, then you should ignore it from the optimising process.

The first thing to look at is your ads – you should be running 5 ads per ad set. If one (or two) are doing significantly worse than the others when looking at stats for the past 30 days, cut them and replace with new ads.

If there are no ads which are significantly behind the pack, then go to the ad set level of your report. Then click the Breakdown menu, and choose Delivery to find a whole list of options to optimise for. I recommend starting from the top and moving down, doing no more than one each time you optimise (and alternating between once of these, and optimising your ads).

Just go through this list, and have a look to see if the CTR of any one of these options is significantly worse than the average for your Ad Set. If something stands out, change your targeting to remove it:

 

Facebook Ads Targeting Breakdown

 

I would recommend keeping a spreadsheet of your changes and making a note of whether your optimisations were successful or not. If you keep up with this, over time your campaign will improve significantly.

Pro Tip: If you want to improve your CTR and your CPC at the same time, then optimise towards your VfM instead, which is your CTR divided by your CPC. You can find out more about VfM and a useful calculator here.

 

How Often Should I Optimise?

Depending on how you spend on your ads, you should optimise every other day, every week, or every month.

A good rule of thumb is:

  • If you spend over £5,00 a day then optimise every other day
  • If you spend over £100 a day (but less than £500) then you should optimise once a week
  • If you spend under £100 a day then you should optimise once per month.

Pro Tip: If you are optimising weekly or monthly, try to do it on the same day of the week or month each time. This will make it easier to compare the effect your optimisations have made when reviewing your spreadsheet of changes.

Whatever you spend, you should give your campaign a week at the start to find its feet. After this you want to give the Facebook Algorithm a chance to catch up to any changes you make, so optimising too often can cause more harm than good.

Basically – make sure your ads aren’t marked as being in the learning phase when you optimise them and you’ll be ok.

 

A Final Warning

You shouldn’t really be optimising towards CTR unless your goal is only to drive people to your page. If there is a different goal to your advertising (such as making a sale) then you should install the Facebook Pixel on your site and optimise towards conversions instead.

A good CTR does not necessarily translate to a good Conversion Rate. If conversions are your goal, then you will supercharge your performance by doing this.

 

[You can find Facebook’s instructions on how to install their pixel here]

 

I hope this helps,

Justin

 

[This originally appeared on Quora]

 

 


 

Q:What techniques are there to get a high CTR on ads on your website?

&

A: “Hi Dilan,

There are lots of things you can do to improve the click-through rate (CTR) for ads on your site.

If you are using an ad network (such as AdSense), then the most effective (and easiest) way to improve performance is optimising where the ads are placed.

The other ways to improve performance are equally simple, even if some of them are counter-intuitive…

[ Before I get started….

Before I go into advice for everyone, a quick note for people who run their own ad campaigns.

Firstly – running your own ad campaigns is only recommended for bigger sites. When your site is small it’s not worth your time to run your own ad campaigns as you won’t make much money (and it’s hard), so just join an Ad Network, follow the below advice, and focus on growing your site.

Secondly – if you are running the ad campaigns which appear on your site yourself then there is too much to explain here – check out this guide instead: Placing Ads On Your Site ]

 

1. Place ads Above The Fold

If you put an ad “above the fold” (meaning it can be seen when someone first loads a webpage) then it will perform much better than ads “below the fold”.

 

What is looks like: Above The FoldClick to enlarge

 

 

This is why Leaderboard ads (728x90s) generally work so well as they are usually placed at the very top of a webpage. You can also place a 970×250 at the top of your site as that generally performs very well, although fewer advertisers make ads in this size, so you might have problems filling it.

Read more about Above the Fold here: Above The Fold Definition

 

2. Choose the right standard ad sizes

The ‘main’ three display ad sizes are 728×90, 300×250, and 300×600. If you put these three ads on your site you will see the best performance. These are the ad units which are used by the majority of advertisers, and so by adding these ad slots to your site you are maximising the amount of money you can make while minimising the amount of effort you have to put in.

Ideally, the 728×90 will appear at the top of the page, the 300×250 will be at least 50% above the fold, and the 300×600 will run down the side of the page.

 

Where to position ads on your siteClick to enlarge

 

This is for Desktop only – for mobile you have an entirely different set of ads to choose from (although the 300×250 is still probably the best performer of the lot).

There are many other ad sizes available, however, if you put odd ad sizes on your site you will find them harder to fill as fewer advertisers use them. Some odd ad sizes perform great anyway, but you have to choose the right ones for you.

Read more about Ad Sizes here: The Ad Size Guide

 

Other types of ads

Pre-Rolls: If you have video’s on your site, then you should add pre-roll ads to them by joining a video ad network. A Pre-Roll is an ad which runs before the video starts – these can be some of the best performing ads in terms of clicks and money.

Overlays: Overlays are a type of pop-up ad which appears within the webpage you are on. These can work extremely well in terms of clicks and money, HOWEVER, they are also extremely annoying. Limiting the number of times they are shown to each user to once per day is vital. By showing the ad less you can maximise the amount of money you make from ads compared to the number of users you lose by being annoying. Also, make sure the sound doesn’t play automatically, and that a close button is easily accessible.

Overlay Example 1

^This is an overlay – also called a Floater in the USA^

 

Pop-up or Pop-Under ads: These are terrible, and only terrible advertisers use them these days – so they don’t pay much. You shouldn’t even consider them.

To find out about other odd types of ad, go here: Non-Standard Ad Units

 

Viewability

To solve the problem of ads loading at the very bottom of webpages (and so never being seen by anyone), the online ad industry came up with a metric called viewability. An ad is considered viewable when it is at least 50% on screen for one second.

 

Viewability Rate FormulaClick to enlarge

 

This might not seem much but after a ton of testing, it has been decided that ads which hit this standard are significantly more valuable (because of their performance) than ads which do not.

It, therefore, follows that you should try and hit this standard for your ads too, at least for the majority of them.

To work out how many of your ads hit this criterion, use this Viewability Rate Calculator: Viewability Rate Calculator

Don’t put too many ads on your site – don’t annoy your users

The most important thing to remember when trying to improve the performance of ads on your site is to not annoy your users. You can mainly do this by not putting too many ads on your site.

Keep in mind, most people will click one ad MAXIMUM when they are on your site. So adding more ads will reduce your CTR simply because of the maths:

 

CTR Formula or Click Through Rate EquationClick to enlarge

 

Basically more ads mean more chances NOT to click rather than more chances to click.

On top of this, you should limit how often annoying ads show – once per person per 24 hours for the most annoying types of ads is best practise. Really though, while you are small you should avoid annoying ad types altogether as they won’t make you much money, but will cost you users. And when your site is small your focus should entirely on getting more users.

Also, keep in mind that ads usually increase the page load speed of a site significantly. The more ads you put on your site, the slower it will be. The slower your site, the fewer people who will visit.

Finally – think about the type of ad you allow. It’s possible that you will make more money by allowing ads for guns, political groups, adult ‘dating’ sites etc on your site, but again you will likely lose users to your site (depending on what your site is about in the first place). Having better quality ads makes people have more faith that your site is a trustworthy place to be – and that trust translates into clicks.

I hope this helps.

Justin

PS These are all ‘white hat’ tips. You can do dodgy things to improve your CTR, but as ad networks don’t want to pay people who cheat the system, they are not worth considering. By setting up ads in a dodgy way you won’t improve your CTR in the long run, you’ll just get yourself banned from ad networks.

 

[This originally appeared on Quora]

 


 

Take care,

Justin