Analyzing Adword Experiment Results

[Sparetoolparts]

Hi Guys

I've ran a campaign wide experiment on Adwords for a month, this was with CPA turned on, on a 50/50 split on my currently/experiment campaign I have spotted some great possible savings on CPC and few high conversions, but tbh I haven't looked at how to properly analyze the results, got any ideas?

I'll post up screen shots later

 

[Sparetoolparts]

Hi guys

Please see some stats:
Clicks
Experiment:179
Original 420

Impressions
Experiment: 5159
Original 15,601

CTR
Experiment: 3.47%
Original 2.69%

AVG CPC
Experiment: £0.04
Original £0.09

Cost
Experiment: £6.69
Original £36.59

Avg Pos
Experiment: 3.6
Original 2.8

Conversions
Experiment: 8
Original 5

Cost / Conv
Experiment: £0.87
Original £7.32

Conv rate
Experiment: 4.47%
Original 1.19%
Unless I'm missing something the experiment is the clear winner it got less clicks, but more conversion and and the cost reduction and cost per conversion is huge. So I would expect if I used the experiment that it should double each value as it was a 50/50 split

 

[Sparetoolparts]

Hi guys

Please see some stats:
Clicks
Experiment: 72
Original 249

Impressions
Experiment: 4070
Original 21,119

CTR
Experiment: 1.77%
Original 1.18%

AVG CPC
Experiment: £0.08
Original £0.20

Cost
Experiment: £5.70
Original £49.76

Avg Pos
Experiment: 3.8
Original 2.5

Conversions
Experiment: 4
Original 4

Cost / Conv
Experiment: £1.42
Original £12.44

Conv rate
Experiment: 5.56%
Original 1.6%

Again big savings and the same conversions, so this would be safe change as well right?

 

[webgeek]

That's the problem with online advertising in general. You start at the bullseye with the best cost per conversion, which doesn't have all the possible conversions, and you want more.

Then, you opt to chase more, at a lower efficiency, hence higher cost per conversion.

Keep chasing outward on the dartboard until you reach the point where you're not getting more conversions at an acceptable cost per conversion. That's your pain point.

This is the point at which people try Bing, A/B test their ads, more clearly define their landing pages (1 page per term per ad group), splitting out their ad groups more clearly (1 page per term per ad group), find additional related terms that might convert, add more negative keywords and demographic segmentation/filters, to improve efficiency so their pain point allows for more conversions to be added (more budget to be spent).

 

[webgeek]

Either way you got 4 conversions - but one way it cost you 12.44 and the other 1.42.

You could do it the new, more efficient way, give me 10 quid for the free advice, and still come out ahead

 

[Sparetoolparts]

Hi guys

Please see some stats:
Clicks
Experiment:226
Original 463

Impressions
Experiment: 6,280
Original 8,684

CTR
Experiment: 3.6%
Original 5.33%

AVG CPC
Experiment: £0.09
Original £0.15

Cost
Experiment: £19.94
Original £64.41

Avg Pos
Experiment: 2.8
Original 1.9

Conversions
Experiment: 14
Original 21

Cost / Conv
Experiment: £1.42
Original £3.21

Conv rate
Experiment: 6.19%
Original 4.55%

This one is different fewer conversions on Experiment but again better costs and cost/conv

Experiment wins again?

 

[webgeek]

Next steps, in all seriousness, is to look at negative keywords to start limiting those erroneous non-buyers, look at what keywords actually converted and prioritise them in bidding slightly, give Bing a whirl and then check your demographics.

Are people converting at a better rate on a particular day of the week, time of day, part of the country, city vs rural, etc. I'd be more inclined, as a gut-guess, to bid up the rural people as they've got less chance of popping into a shop and buying a Bosch drill chuck, or hammer drill spindle, when they need one, without driving 100 miles. But that's just a WAG.

 

[webgeek]

Almost always I've found a better conversion rate at position 3-10 (old school, when listings were top, side, bottom). You get much more impulse clicks at position 1. Great if you're going for max conversions - bad if you're going for max efficiency.

If you can convert at position 2.8, as you did here, or 3.8 as in the first experiment, then the best ROI would be from running more more more product ads, lower down the page, closer to position 5. That's where you find considered purchases from people who shopped the competition and who didn't buy.

 

[Sparetoolparts]

Yes my next experiments will be based on time of the day, as I would suspect most sales are coming from companies so 9-5 hours may need more budget thrown at them, before I experiment again I want to make sure my first round of decisions were done correctly, I've never actually looked at this b4 so I want to get it right

 

[Sparetoolparts]

Hi guys

Please see some stats:
Clicks
Experiment:510
Original 459

Impressions
Experiment: 6,409
Original 5,870

CTR
Experiment: 7.96%
Original 7.82%

AVG CPC
Experiment: £0.11
Original £0.11

Cost
Experiment: £57.79
Original £49

Avg Pos
Experiment: 2.3
Original 2.5

Conversions
Experiment: 33
Original 16

Cost / Conv
Experiment: £1.75
Original £3.06

Conv rate
Experiment: 6.47%
Original 3.49%

Ok I would say another clear win for experiment it did cost more but the conversions doubled and conversion rate is double too

 

[Sparetoolparts]

Hi guys

Please see some stats:
Clicks
Experiment: 836
Original 1,104

Impressions
Experiment: 18,654
Original 28,297

CTR
Experiment: 4.48%
Original 3.90%

AVG CPC
Experiment: £0.06
Original £0.10

Cost
Experiment: £51.41
Original £108.74

Avg Pos
Experiment: 3.1
Original 2.2

Conversions
Experiment: 18
Original 24

Cost / Conv
Experiment: £2.86
Original £4.53

Conv rate
Experiment: 2.15%
Original 2.17%

I would still say Experiment won, as although fewer conversions the cost saving is impressive

Experiment wins again?

 

[Sparetoolparts]

so that's me done, please let me know your thoughts on the winners out of these campaigns as I will be applying the changes tomorrow

 

[webgeek]

Given the number of items in your store catalogue - I'd be thinking all the experiments were solid wins. Anything that drives down your conversion cost gives you room for another product to be advertised and profits to increase.

You doing straight Adwords, or doing some Product Listing Ads?

I've almost always found Bing to have much lower volume and much higher conversion rates. If you're not doing so already, it might be worth a shot.

 

[StevePoster]

Its really a benefit to be a member on this reputable forum site. Anyway, @webgeek is there any studies or proven data that Bing provides higher conversion rates despite of its lower volume?

 

[garyk]

One interesting factor is that the demographics for Bing favour older users (dare I say it silver surfers! )

Reason being an older person buying a windows PC is far less likely to install another browser and will use IE and of course its default search engine; Bing.