Making Decisions Based on Conversion Data

Changing your PPC spending based on conversion data and/or Web Analytics data is going to require the application of a lot of intuition and wisdom. It is NOT as simple as seeing lack of ROI and making a change.

I'll post some specific real-life examples in this thread to illustrate some of the thinking you'll want to apply.

OBSERVATION: Your cost per conversion on some general keywords is right around $100 - too high given that your average profit per sale is $60.

DECISION: Reduce bids to lower that cost per conversion to $60.

UNEXPECTED RESULT: Phone orders show an enormous drop.

REACTION: Raising those bids again restores phone orders.

CONCLUSION: General keywords are more likely to generate phone orders. (Those who are better at using the Internet are generally better at finding what they want and buying online than those who do general searches, need help finding what they want, or aren't comfortable buying online.)

Another example. You sell a range of products from under $100 to $5,000 or more. Conversion data indicates almost no conversions from ad spend on your most expensive products.

Before you pause that campaign be sure to check phone orders. You are likely to find that almost all orders for that product line are placed over the phone - NOT on your site. This is especially true if there are many options or decisions to be made when ordering.

It is very likely your ads are driving those sales. Turn them off due to lack of online conversions and you could seriously impact revenue and profits.

IMPORTANT POINT: The person making advertising decisions must have input about offline sales, consider that information in the decision making process, AND receive feedback on any changes made.

Encourage sales to notify advertising of any major changes. You may even want to discuss planned changes in advance or at least notify sales that a change has been made and for them to let you know how it affects them.

Yes - all is not as cut & dry as it may seem - there is "a lot of intuition and wisdom" necessary when determining what works and what does not. A holistic approach to all your marketing on & off-line should be taken.

If your marketing team is not conversing with your sales team - who is not conversing with your design team - who is not conversing with your business administration team - who is not conversing with your management team - you will not meet your marketing potential no matter how large or small you are.

When one brings it all together and makes the best advantage of all efforts in a synergistic environment then good things can happen.

I would like to also point out that a company that hires a "seo or sem" expert or "internet consultant" that DOES NOT give them full communication and liaison with their other experts (in regards to off-line marketing etc.) it is like strangling your own foot and then wondering why you cannot walk.

Here is another example for those making changes to increase their ROI. Be cautious of pausing low traffic keywords too soon. Some of your highest profits will be driven by keyword phrases that only generate one sale a year - sometimes from only one click.

Don't assume that only those keyword phrases that converted during any short period of time will bring you sales. Even your higher traffic best performing keywords may not generate sales in any particular week or month.

{looks around - errm sorry, did I interupt your conversation?}

Does anyone have any suggestions they can share on how they use conversion tracking or analytics to increase their ROI?

If you have questions or would like to start some discussion on any specific data or decisions feel free to post here or in a new thread.

I can vouch for this.

Having just been through the same kind of thing with our campaigns where analysis was done and it was decided to chop all of the non converting or poorly converting keywords.

This was against our better judgement but we were under orders from high up to do it based on figures on a piece of paper.

It had a negative impact on the campaign as a whole and we have ended up putting back some of the phrases that were cut out.

The cutting of the keywords (usually generic keywords with high traffic but little or no conversions) has resulted in terms that used to convert very well being impacted.

It leads us to the conclusion that generic terms (one or two word phrases usually) drive conversions to other keywords that are very profitable or they drive sales to organic listings.

Take Rose's advice, think very very carefully before making the decision to delete or severely alter your keywords, Adwords is not an exact science and cannot be driven by numbers on a piece of paper.

We are also investigating how to track these campaigns more effectively to try and make a bit more sense out of it all. If we make any progress I'll report back.

Thank you for sharing your experience with everyone. In further researching how Google Analytics works it is obvious to me why this data is so inaccurate. Check out the thread at What Source Really Generated That Conversion? for details.

I believe that potential buyers first search on very general keywords to get an idea what they will later buy. Then almost every way they come back to purchase overwrites the original PPC ad tracking code from the general keywords so they show almost no conversions even though they are how eventual buyers first found the seller.

This is due to Google Analytics (GA) and many other Analytics programs using the last click rather than the first click as the source of conversions and revenue. I am researching into whether it is possible to reconfigure GA to use first visit rather than last. That one change would make an enormous difference and I believe provide what we really need to manage PPC.

Without that first click leading to awareness of products and a particular seller many of those last clicks would never occur. It is the first click that is actually driving the eventual sales - not the last one.


[ Message was edited by: flyingrose 10/31/2006 03:09 pm ]

Rose.

That is what we have found.

We have also found that CPC campaigns drive traffic to shopping sites and to a lesser extent affiliates.

It's our unconfirmed belief that you can be paying 3 times for the same conversion before you actually convert. This is because people by nature will always look for a bargain, be it either on Auction sites, shopping comparison sites or affiliate links for voucher codes and special promotions.

We are working on a way of trying to manipulate the CPC links to better track the conversions through our own site analysis program but so far we are hitting a dead end.

We are tailoring our campaigns now on a profit per keyword scenario and we are structuring our campaigns accordingly, we are still in the early stages of this and I will report back any findings we have, initial signs look good as we have for the first time ever managed to convert an entire campaign at over 3% for a whole week (Bear in mind that this is for an average transaction value of over $400).

On the back of this we have constructed a model whereby we can take all cost factors in account for a product (Marketing costs, Shipping Costs and other Overheads) and analyse pricing from this.

It's a small success that we are hoping to build on and expand to other campaigns.

Thank you for the additional detail. I can confirm that increasing PPC ads definitely drives sales that appear to originate on shopping sites and pricee comparison sites.

It is highly likely that someone will check your prices against those types of sites. If you also have products listed there they are most likely to come back to your site through a tagged link which I am told overwrites the PPC tag.

Most sales incur multiple click charges even if you only use PPC and nothing else. Many searchers will look at other sites and come back to your site from your Google ads. Google does charge for multiple clicks and considers them valid clicks. Not too painful when bids are low but progressively more expensive as bids go up.

They are also highly likely to buy later than they are to buy immediately and many will come back the same way the next time - through a PPC ad click. (This is the primary reason I highly recommend against limiting ads with budgets. Any ad that runs should appear consistently. If advertising spend is too high better to pause some ads or lower bids.)

One point you may wish to make regarding trying to make every single product's advertising profitable is the likelihood of a person buying multiple products and placing repeat orders.

When first buying from a new site some buyers will start with an inexpensive item to make sure the company is legitimate and the service is good. Then they'll make progressively larger purchases.

Trying to pinch the pennies too hard may be far more costly in the long run than can ever be measured.

Here are links to additional information on this subject:

What Source Really Generated That Sale?

The sticky thread Using Web Analytics including Google Analytics at the top of this PPC forum is specifically about Analytics. That thread has links to many of the best Analytics blogs and other resources.