Making money with search: How conversion really works!
I haven’t blogged for quite some time, but I was surprised the following information was still an eye-opener for so many people. So here an eye-opener on how conversion and search engine marketing really work. How do you use SEA and SEO to its full potential and how do you make money with SEM?!
Investment per conversion
From ranking on a search term to a closed deal, a conversion undergoes several stages. These stages are all as important as the investment for achieving a ranking. They are the real money making factors in search engine marketing.
$ ≈ # ≈ ☺ ≈ $
Investment per conversion can be seen as the formula above. It states that:
- $ ≈ # The investment for ranking:
What is needed to rank for a particular search term? This factor is mainly determined by the competition (both SEO and SEA). - # ≈ ☺ The chance of being clicked:
Is my result relevant enough for that visitor? Getting many clicks on lower rankings can easily beat 1st positions. - ☺ ≈ $ The chance of a converting visitor:
Do I offer the right service for the visitors desires and is there an easy navigation? Even small improvements in usability can double your conversions. - $ …≈… $ The investment and profit per conversion:
What is a conversion worth and do the costs match? Having a bigger margin creates bigger budgets to outrank the competition.
The ≈ is used to state that there is uncertainty. It equals the chance of something resulting in a desired action. Your SEM knowledge can make that chance far easier to determine in advance.
Calculation example
To clarify how important conversion is I will give two Google Adwords examples taken from real websites. These are Dutch examples in the credit business. I’ve converted the costs-per-click using the current dollar to euro ratio, and simplified some figures for easier calculation. The biggest difference between Website A and B is that B uses its homepage as landing page and A creates specific landing pages for every single search term.
| Website A | Website B | Comment | |
|---|---|---|---|
| search term | doorlopend krediet | doorlopend krediet | Dutch credit term |
| max CPC | $ 10.50 | $ 7.00 | B isn’t willing to bid more |
| avg position | 2 | 7 | Browsing visitors will more likely click lower results as well, but are less likely to convert directly |
| CTR | 10% | 1% | Of 150 searches per day |
|
When we take 100 visitors the average percentages from the landing pages are as follows. |
|||
| Visiting order form | 70% | 50% | Percentage of visitors from this search term |
| Finishing order form | 30% | 6% | Percentage of visitors that started order form |
| Total percentage of visitors | 21% | 3% | From visitor to finished order form (=conversion) |
|
This comes down to 21 and 3 visitors of every 100 visitors convert. The cost per conversions then comes down to: |
|||
| 100 visitors | $ 1050 | $ 700 | To buy from Adwords |
| conversions/100 visitors | 21 | 3 | These filled in the form |
| CPA | $ 50 | $ 233 | Amount/conversions |
|
When we see that the profit per conversion isn’t that different, Website A is doing OK, but B needs to increase its conversion rate. |
|||
| Profit per conversion | $ 64 | $ 70 | Case A uses more manual labour to convert a lead to a loan. |
| Profit-CPA | $ 14 | -$ 163 | Not measuring CPA ruins Website B. |
What can we conclude?
- Website A is willing to pay more per visitor, but they measure better if they convert. They know their real maximum.
- The percentage of people starting the order form is enormous, but the obvious button on a very relevant landing page (doorlopend krediet landing page) gives Website A an even better CTR from the landing page than Website B has from the homepage (their landing page) (70% vs 50%).
- The order form in Website B has 5 steps across 5 pages and they require you to fill in everything. Website B uses 3 steps on 1 page (loan request form) and requests any missing information using a call-center. This is more expensive, but converts at least just as good. Therefore a lower profit per lead.
This is a real example and I know these percentages seem extreme, but I assure you these differences are common in many industries.
Website A can do much better!
Website A seems to have everything well in order, but they can use search engine marketing even smarter. The credit industry is extremely competitive, but other less competitive industries might offer chances for loans. Do you see the relevance between a divorce and loan? Well, visitors sure do.
Website A created a section specifically about financial problems that arise with a divorce. You suddenly need a lawyer, an extra home, extra furniture and you might need to pay alimony. All costs that come very inconvenient during a period that is tough enough by itself. A loan might offer some comfort to get things in order.
After introducing this new section, Website A did multi-variate-testing on Ads and landing pages. Below we compare the original situation to the situation that was achieved after testing.
| Original situation | After testing | Comment | |
|---|---|---|---|
| search term | scheiding aanvragen | scheiding aanvragen | Dutch for “request divorce” |
| max CPC | $ 1.00 | $ 1.00 | Same bid |
| CTR | 2% | 8% | 4 ad versions tested. This also means we get more visitors from this search term. |
| avg position | 3 | 3 | Same position |
| actual CPC | $ 0.89 | $ 0.70 | CTR Quality score decreases CPC |
|
When we take 100 visitors the average percentages from the landing pages are as follows. |
|||
| Visiting order form | 35% | 50% | 4 landing pages tested. Divorces are less relevant, so this is worse than loan terms |
| Finishing order form | 20% | 24% | Better expectations from landing page increase form finishes |
| Total percentage of visitors | 7% | 12% | from visitor to finished order form (=conversion) |
|
This comes down to 7 and 12 visitors of every 100 visitors convert. The cost per conversions then comes down to: |
|||
| 100 visitors | $ 89 | $ 70 | to buy from Adwords |
| conversions/100 visitors | 7 | 12 | these filled in the form |
| CPA | $ 12.71 | $ 5.83 | amount/conversions |
|
The profit per conversion from the previous example still applies but compared to “doorlopend krediet”, “scheiding aanvragen” is much cheaper. |
|||
| Profit per conversion | $ 64 | $ 64 | Case A uses more manual labour to convert a lead to a loan. |
| Profit-CPA | $ 51.29 | $ 58.17 | After one month of multi-variate-testing. |
Compared to “doorlopend krediet” from the first example, “scheiding aanvragen” creates a far larger profit. The amount of searches for “scheiding aanvragen” is lower than “doorlopend krediet” and therefore both are needed to fill the demand for loan leads. But every less competitive industry we can associate with loans can be added to our audience. Combined these have a great potential and the profit is enormous “until other banks discover them”. So I’m not going into the details, but you should get my point. When you keep your Profit per conversion high and cost per conversion low, you can outbid any other company (especially from other industries).
June 30th, 2008 at 1:37 pm
Nice to read. In particular the examples given. And although this is almost the basic of webmarketing, its always nice to read how to calculate properly the ROI of marketing investments because it stresses the importance of it!
June 30th, 2008 at 3:17 pm
Nice information. I have noticed the same thing with a homepage as landing page.
Keep posting Peter!
July 7th, 2008 at 11:05 pm
Hey Nice article;
For fun i tested the nr.1 position for “lenen” (loan in english). I placed a 25 euro bid (first a 20 euro bid) and got concequently on the nr.1 spot, above frisia.
Like a grace from heaven, only 1 person clicked.. which costs me 19,11 euro (yes about 30$).. The F*cker didn’t convert; but i do have a screenshot of the number 1 position:
://omblog.nl/de-nr-1-yes!!.jpg
Although it was for fun; it did feel good :)
Obtaining the lowest CPC is the most important thing on my mind, so right now i’m experimenting with it..
October 17th, 2008 at 12:52 pm
Thank you for the excellent post, the examples are very clarifying! I now have a much better understanding of how conversion works.
I can’t wait for your article explaining how to effectively run a split test, that’s another thing I would love to learn more about :)