Advert Text Blog Posts

The Hidden Cost Of Increasing Your Bids To Move Up The Search Results

Thursday, August 30th, 2007

Advert Text, Bidding, Google Adwords

I wrote the other week about how to find the sweet spot for your advert, the position that generates the most profit for you. But it’s a long and slow process, if you are just interested in making a small change to your bids (maybe as a result of a change to your Quality Score or conversion rate). There’s a shortcut that you can take, but only if you’re careful…

Suppose you’ve been paying £0.50 per click, and making some money, you figure that since a click is worth £1 to you, increasing your bid to £0.70 would be worthwhile. So, you increase your bids, you move from 5th to 4th in the results, and get 120 clicks instead of 100. Job done, and all is well with the world. Erm… no… You are now paying £84 instead of £50, and getting 120 clicks instead of 100. That’s £34 more, for 20 more clicks, or £1.70 for each additional click! Which is more than they’re worth. The extra clicks each cost you £0.70, but every other click (that you were going to get anyway) has cost you more, and so you have to include this extra spend in your calculations…

The only way to use this sort of approach is to look at the cost per additional click, rather than the overall cost per click. When the cost per additional click reaches your breakeven point, you have reached your sweet spot – this is, in fact another way of calculating the sweet spot. Right then, example time…

Take the above example. Clearly, since the additional clicks would cost £1.70 each, you don’t want to do that. In fact, it was such a bad idea, you probably want to go the other way…

You reduce your CPC from £0.50 to £0.45, and the clicks falls from 100 to 95 per day. Instead of spending (£0.50*100) = £50, you’re now spending (£0.45*95) = £42.75. A quick sum tells you that the clicks you just ditched were costing £7.25 for 5, or £1.45 each. Too expensive, so you should move your advert down.

So you try £0.40, and now have 85 clicks. Another quick sum, and you can see that the clicks you have got rid of were costing £0.88. You were making money on these, so you should keep them. And so you have your sweet spot.

To prove it, here’s the profit you would be making in each position:

CPC Clicks Cost Profit
£0.70 120 £84 (£120-£84) = £36
£0.60 110 £66 (£110-£66) = £44
£0.50 100 £50 (£100-£50) = £50
£0.45 95 £43 (£95-£43) = £52
£0.40 85 £34 (£85-£34) = £51
£0.35 70 £25 (£70-£25) = £45

This is a quick way of telling whether you are in your sweet spot, and whether the change that you just made was a good one or not… Since you can see on your campaign summary how much you spent, and how many clicks you got over a specified period, so you can certainly get a quick and easy answer to whether your latest change worked, without having to do lots of sums…

no comments

Why Making Money On Competitive Terms Is No Harder Than On Uncompetitive Terms

Wednesday, July 18th, 2007

Advert Text, Bidding, Dynamic Keyword Insertion, Google Adwords

Seriously, if you are using the ‘sweet-spot’ approach to decide what position your advert should appear in, the amount that everyone else is bidding makes no difference to whether your campaign makes money. It does, however, affect how much money you can make. Not convinced? Look at the following example. Take the following two search terms. The profit from a sale, and the total traffic is the same on both. So are the clickthrough rates in different positions and the conversion rate. The only thing that’s different is the cost per click of appearing in each position.

Position Cost/Click Clickthru Rate Conv. Rate Impressions Clicks Conversions Profit
1 £1.00 15% 6% 1500 225 14 £45.00
2 £0.90 12% 5% 1200 144 7 £14.40
3 £0.80 10% 5% 1000 100 5 £20.00
4 £0.70 9% 5% 900 81 4 £24.30
5 £0.60 8% 5% 800 64 3 £25.60
6 £0.55 7% 5% 700 49 2 £22.05
7 £0.50 6% 5% 600 36 2 £18.00
8 £0.45 5% 5% 500 25 1 £13.75
9 £0.40 4% 5% 400 16 1 £9.60
10 £0.35 3% 5% 300 9 0 £5.85
Position Cost/Click Clickthru Rate Conv. Rate Impressions Clicks Conversions Profit
1 £2.00 15% 6% 1500 225 14 -£225
2 £1.80 12% 5% 1200 144 7 -£115
3 £1.60 10% 5% 1000 100 5 -£60
4 £1.40 9% 5% 900 81 4 -£32
5 £1.20 8% 5% 800 64 3 -£13
6 £1.00 7% 5% 700 49 2 £0
7 £0.80 6% 5% 600 36 2 £7
8 £0.60 5% 5% 500 25 1 £10
9 £0.50 4% 5% 400 16 1 £8
10 £0.40 3% 5% 300 9 0 £5

In the first case, the sweet spot is in fifth place (profit = £26), and in the second case, it’s eighth (profit = £10). But in both cases, the Cost Per Click is the same. This is because the optimal cost per click has nothing whatsoever to do with the number of clicks or conversions that you get – it’s purely dependent on the conversion rate, cost per click and the profit (exc. the ad cost) you make on a sale. And these things are totally independent of where your advert appears in the search results. It’s not that surprising really, if you think about it. If your cost per conversion is lower that your profit per conversion, you make money – otherwise you don’t. And cost per conversion can be expressed as cpc/conversion rate. Since the conversion rate is (usually) unaffected by result position, then a cpc will be profitable or not irrespective of your position in the results. One interesting implication of this is that it answers the question – “what should I do if somebody jumps above me in the search results?”. The answer is of course (!) that you should do nothing, and accept the reduced profit (my blog on the numpties makes a lot of sense, doesn’t it!). In summary, ignore what everyone else is doing – work out the right bid for you, and stick to it.

no comments

A:B Advert Testing, A Cautionary Tale

Monday, July 9th, 2007

Adgroups, Advert Text, Google Adwords, Pay Per Action, Testing

The conventional wisdom on PPC adverts on Google is that you should look to improve the click through rate, as it is generally accepted that this is an important attribute in the Quality Score, which determines the amount that you need to bid to get a certain position (or how high up the rankings you appear for your bid, if you prefer). And this is probably true, and isn’t a bad idea. But it’s definitely not a good idea to focus on the click through rate to the exclusion of all else. The click through rate is an indication of how interested people are in your advert, but if your advert does not accurately represent the content of your site, you’ll be enticing traffic that doesn’t convert very well, and may be putting off exactly the people that you should be attracting to your website. This sounds like an easy thing to avoid, but it’s not quite as straightforward as it sounds. Suppose that you are a company that offers free marketing advice via a weekly e-mail that people have to sign up for. Your initial advert may read:

Free Marketing Advice Get Free Advice From Marketers Inc Free E-Mails Every Week MarketersInc.com/Advice

The advert does quite well, and gets conversions occasionally. But you’re concerned that the second line is a fairly weak call to action, so you decide to try something different.

Free Marketing Advice Get Free Advice Here! Free E-Mails Every Week MarketersInc.com/Advice

You run it for a while, and it doubles the click through rate, so within a day or two you bin the old advert and go forward with the new one. Then you look at the third line. It doesn’t really extol the benefits of the e-mails, so you try another line.

Free Marketing Advice Get Free Advice Here! Learn The Tricks Of The Trade MarketersInc.com/Advice

Even better click through rates, so you keep this one. But the changes in the second line may lead people to believe that there is free information on your website, rather than from a marketing company. Whilst you’ll get more traffic to your site, it’ll be of poorer quality. And the change to the third line reinforces this. But surely you’ll see a fall-off in the conversion rates, and keep the old adverts? Not if you’re changing your adverts as soon as one appears significantly better than the other, based on click through rates. Suppose that the campaign above starts out with a click through rate of 3%, then increases to 6% and 8%. At the same time, the conversion rate falls from 10% to 7% to 4%. Finally, assume that the cost per click moves from £0.30 to £0.28 to £0.25 If you accept a 90% level of significance, your results look something like this. TABLE 29 There is no real falloff in the number of conversions, and a significance test of the difference in conversion rates is totally insignificant. In fact, to get significant results (even at the 90% level) for the conversion rates, you’d need to wait much longer. Table 30 To put that in context, if you were getting 400 impressions per day, the tests for click through rates would take (1 + 3 =) 4 days, whereas the tests for conversion rates would take (38 + 8 =) 46 days. That’s quite a lot longer. So, what’s the conclusion here? Should you run your campaigns for ten times as long, to confirm that the new advert doesn’t hit your conversion rates? Bear in mind that the changes above are quite extreme , it’s unlikely that your results will show anything after waiting ten times as long , when do you draw the line, and say that the change is too small to matter? Even here, we’ve not taken into account the impact of reducing the cost per click (which will slightly offset a reduced conversion rate), or the impact of increasing the total number of conversions (even at a slightly higher cost per conversion, this could still be a good thing). Alternatively, should you just ignore the conversion rate, and hope for the best? Or try very hard not to change the meaning of the advert? You only need to write one bad advert to wreck your campaign. Perhaps the best approach is to physically look at the conversion rates of the adverts that you are dropping , if they are lower, then ask the question œhave I caused this to happen? The fewer conversions that you are getting, the harder it’ll be to stop a problem , so monitor the conversion rate, and if it starts to drop, check to see if you’re the cause.

no comments

A:B Advert Testing – Is Statistical Significance Over-Rated?

Friday, June 29th, 2007

Adgroups, Advert Text, Google Adwords, Pay Per Action, Testing

On the face of it, probably a bit of a daft question. How can you be sure that your new advert is better than the old one, if you don’t wait to see if it’s statistically significant? And to an extent, that’s true. If you were to ignore significance completely, the moment somebody clicked through one of your adverts, you’d decide that it was the better advert, and bin the other one. It’s quite possible that only 50% of the time you’d select the better advert, and for every improvement that you make to your advert, you make another change for the worse, and you don’t get any overall improvement at all. But there’s a trade-off for statistical significance. Suppose that you have two adverts, one that generates a click-through rate of 5%, and one that generates a click-through rate of 10%. How long should you wait before you are sure the 10% advert really is better? If you get 30 impressions per day, it’ll take four days to be 85% certain (3/60 vs. 6/60 is significant at the 85% level). But if you want to be 95% certain, it’ll take eleven days (8.25/165 vs. 16.5/165 is significant at the 95% level). And to be 99% certain, it’ll take twenty days! So, in the time that it takes to run one test at the 99% level, you can run five tests at the 85% level. Clearly, you can get far quicker improvements in your overall click-through rate, if most of these changes are genuinely for the better. But what about the risks? You could select to keep adverts that are, in fact, worse than the existing ones (and you will, 15% of the time , any change to an advert will change the click-through rate; there are no ˜equally good’ adverts). But I would challenge that if an advert appears better at the 85% level, whilst it may be worse, the chances are very small that it’ll be much worse. So, if you run five tests in those twenty days, you’ll probably make one change for the (slightly) worse, and four changes for the better. Still an improvement on the one change that you’d make if you were determined to wait until you were 99% certain that you were making the right choice , this is advertising, not a clinical trial! Of course, this is a bit of an over-simplification. In reality, most of your advert tests will yield a much smaller return than doubling the click-through rate, and a lot of them will not be better than the old advert. The first point here is quite important , the smaller the difference between the two adverts (increasingly true once you’ve entered an ongoing process of testing), the longer it’ll take to get strong significance, and the less risk there is in taking the wrong option occasionally. For example, if you were getting 30 impressions per day, and had adverts with 5% and 6% click-throughs, you’d get 85% significance after 75 days, but even 95% significance is going to take 193 days , nearly three times as long. As for the second point, what if the new advert is performing worse than the existing one after a few days? It’s not significant, but, in a mirror of the argument so far, if it is in reality a better advert, is it likely to be much better? Is it worth waiting weeks to see if this advert, that’s probably worse than the existing one, is actually slightly better (remember that the smaller the difference, the longer it’ll take to be sure). Perhaps the time is better spent writing a new challenger, which may prove itself quickly? So what level of significance should you use? Personally, I’d say that 85% is probably sufficient, but I can see an argument for 90%. I feel that running a test for three times as long (as an 85% test) to get to 95% is excessive , yes, you’ll get it wrong less often, but it’ll take a lot longer to generate improvements, and lets face it, your rivals probably aren’t standing still! There is, of course, one problem that brings the whole process to a grinding halt. What if the two adverts are producing very similar results? It’s widely acknowledged that a small change to an advert can have a big impact, but more often than not, it has a very small impact. Everything stops until you get significant results, and the more similar the performance of the adverts, the longer it’ll take. The solution is fairly clear , sooner or later, you’ll have to stop the test. You can either keep the existing advert, since the new advert hasn’t proven itself, or you can take whichever is the better to date, regardless of whether it’s significant or not (this’ll be the better advert more often than not). I’d advocate the second option, although really, it doesn’t make much difference which you choose (since they are performing very similarly). An interesting claim , that under certain circumstances, you should take the advert that is performing better, regardless of whether it’s significant or not! So what process have we arrived at?

  1. Decide before you run your new advert how long you are willing to wait for a result , this’ll depend on how long you’ve been testing (as you go on, the chances of finding a quick, big win decrease) and (obviously) how many impressions you are getting.
  2. Set the advert live, checking regularly for significance. I’d recommend www.splittester.com, but any testing tool will do.
  3. If, after a few days (longer if you’ve got little traffic), the new advert is worse than the old one, kill it, and write a new advert.
  4. Once you’ve got 85% significance (or 90%, if you’re of a nervous disposition), keep the better advert.
  5. If the deadline set in step one is reached without a significant result, keep the better advert, regardless of how small the difference is.

1 comment

Multivariate Testing On PPC Adverts

Friday, June 29th, 2007

Advert Text, Google Adwords, Testing

When most experts talk about testing adverts on Google Adwords, they recommend that you change only one thing in the advert, so that your results aren’t confused by conflicting changes (if one change improves your click through rate, and one makes it worse, then the overall impact could be better or worse, and you’ve no idea what worked at what didn’t).

And that’s all ducky. But it assumes that the lines of your advert work independently of each other. What happens if this isn’t the case.

Consider the two adverts below:

A)
ABC Mortgages
Free Advice And Valuations
Competitive Rates

Suppose that you’ve been running A) for a while, and try a new third line œQuick And Friendly Service:

B)
ABC Mortgages
Free Advice And Valuations
Quick And Friendly Service

You get the following results:

A) 10000 impressions, 1000 clicks, click through rate = 10%
B) 10000 impressions, 800 clicks, click through rate = 8%

So you keep advert A)

Then you try it against another advert, with a second line of œYour Local Mortgage Broker:

C)
ABC Mortgages
Your Local Mortgage Broker
Competitive Rates

You get the following results:

A) 10000 impressions, 1000 clicks, click through rate = 10%
C) 10000 impressions, 800 clicks, click through rate = 8%

Again, you keep advert A).

But have you missed something here? We’ve no idea if advert A) is better than an advert with both changes:

D)
ABC Mortgages
Your Local Mortgage Broker
Quick And Friendly Service

It’s quite possible that the two new lines work well together, as they have a similar tone. This being the case, you could potentially get the following results:

A) 10000 impressions, 1000 clicks, click through rate 10%
D) 10000 impressions, 1200 clicks, click through rate 12%

This is quite possible, but you can’t run this test instead of the others, in case one of the lines was better, and the other worse. And if you run it as well as the others, it feels a bit like you’re wasting time. And if you wanted to try varying all three lines, you’d need to run seven different tests, instead of just three.

In principle, the solution is quite simple. There’s an easy way to test your three adverts in the time of two tests, or if you are changing all three lines, eight tests in the time it would take to run four tests:

Simultaneously, run all possible combinations of the adverts against each other at once.

The above tests ran for a total of 20,000 impressions each, but how many impressions would it take to get the best of the four options?

Only 40,000 impressions, because you don’t need to waste 50% of your impressions on the control, only 25%. This is the same amount of time that you’d have spent just testing A against B and A against C.

So your results would have looked something like:

A) 10000 impressions, 1000 clicks, click through rate 10%
B) 10000 impressions, 800 clicks, click through rate 8%
C) 10000 impressions, 800 clicks, click through rate 8%
D) 10000 impressions, 1200 clicks, click through rate 12%

In fact, you can save a bit more time by ditching one of the adverts early on, if it’s clearly not going to be the best.

So perhaps testing one change at a time isn’t necessarily the best idea , the stronger the synergy between the lines of your advert, the more likely you are to miss positive changes.

However, there is a price to pay for this approach (isn’t there always?). The test is slightly more likely to come up with the wrong answer (it’s slightly less significant), so you’ll probably want to run it a little bit longer to get the same degree of confidence in your results.

In summary, if you think there’s likely to be a strong synergy between two lines of your advert, it’s worth testing all four combinations of the two lines at the same time. If there’s no reason to believe that the two new lines work particularly well together, but poorly individually, then there’s no benefit to running the four options at once, and you should run two a:b tests as usual.

For example, if you have the advert:

Dave’s Confectionary
Buy Our Chocolates!
Free Next Day Delivery

And you want to try a second line of œGet Luxury Chocolates Here!, and/or a third line of œGift Wrapping Available, you have to decide whether these lines are likely to complement each other, but are unremarkable individually. Here, it looks like the lines are pretty much unrelated, with one being an alternative call to action, and the other suggesting a different feature. So two standard (one change) a:b tests would be quicker or more reliable (depending on how long you wait for your results) than running all four possibilities together.

Any comments? Disagree? Let me know¦

no comments

« Older Entries
Newer Entries »