Low Quality High Volume Clicks
When I launched my first pay-per-click campaign I was seduced by the thought of getting as many clicks as possible. I figured that with a lot of clicks I would get more sales and sales would be a percentage of all the clicks I received. Unfortunately, the conversion, that is the ratio of sales to clicks, where consistently much lower than 1%.
Where I went wrong was that I estimated an advertising budget and divided it by the profit margin of the product. This gives the number of sales that I would need to make to cover the advertising budget. I estimated that my conversion rate would be around 1-4%. The basic math is shown below:
- Advertising budget := $240 per month
- Estimated Product Profit Margin = $8.00
- Sales to cover advertising budget = $240/$8 = 30 units per month
- Estimated conversion rate: 1-4% — assume around 2.5% on average
- Clicks needed to make 30 sales: 30/0.025 = 1200 clicks monthly
- Maximum bid price per click $240/1200 = ~$0.20 maximum bid per click.
The problem with this model is that the lower your conversion rate the more clicks you need to make the required conversions. And the more clicks you need, the lower the maximum bid price.
Unfortunately, as in everything else, the quality of the click, in this case measure by how likely it is to be converted to a sale, is directly related to the the maximum price. And if you follow this model what happens that you get a lot of low quality clicks with very few sales. This causes your advertising costs to exceed sales revenue. At least, this is what I experienced. Therefore, I needed to modify my strategy.
Strategy 2.0 — Keyword Search
What I tried next was to very carefully select my targeted market. The pay-per-click service I am using has a tool that provides keyword suggestions and I used it to explore my options.
I found that some of the keywords had first page bid estimates that exceeded the estimated profit margin. I choose not to compete for those keywords and excluded them from my list.
On the bids that remained, the ones I could competitively afford, I initially bid at the estimated maximum. As a result the conversion rate significantly increased to well above 2% on the average. Furthermore, the costs per conversion decreased. This is the advertising costs for acquiring a sale.
The PPC service I was using provided these numbers as part of the service. I had to set up the tracking which required that I load a custom script (provided by the PPC service) on the Web page where the customer confirmed a sale.
I would highly advise setting up the scripts and PPC service to monitor costs and performance. This information is required to effective manage the PPC campaign.
Strategy 3.0 — Content
At first, I had set bid for the content ads (image ads that are placed on affiliated network sites) at a relatively low amount. However, when the number of clicks I was receiving plummeted I panicked and I reset the PPC services to automatically manage the maximum bids costs. This had an immediate effect in that the number of clicks skyrocketed and so did the associated advertising costs. Unfortunately sales did not come even close to tracking the costs. So I reset the content bids amounts back to manual management and relatively modest bid amounts.
For content advertising, carefully select the relevant placement networks. Also be aware that its is widely reported that content PPC is prone to click fraud. I have not experienced that myself but there are numerous mentions on the Web about this subject.
Another thing to be aware of is that content advertising is not as targeted as keyword search advertising. That is, in a keyword search your ad is placed in front of someone that has expressed an interest in the product — at least that should be case if the keywords have been selected properly. Whereas, in a content placement, the ad is placed in a broad category of similar products. As a result, the conversion rates will, in general, be significantly lower than keyword search placed ads.
Strategy 4.0 — Weed the Garden
Once a few weeks of performance data has been collected on the PPC campaign carefully analyze the costs and the associate revenue of keywords, content networks, text and image ads. Remove those elements that costs more than they produce in revenue. Next examine those elements that remain and try to identify common traits. Use these common traits to add new elements with comparable traits.
Furthermore, use your intuition to select completely new untried elements. In short, experiment with something completely new and untried.
In addition to eliminating advertising elements that cost more than they produce, look at the days of the week that the ads run. If there are any days where the daily advertising budget exceeds the revenue consider not running the PPC campaign during those days.
A word on caution on this point, the decision to purchase may lag when the ad was seen by some considerable time. That is, the customer may see the ad on Friday and make the purchase on Monday. So if there is a sudden drop in sales when the PPC ad is shut off on select days then things need to be reconsidered.
Also, if the advertising costs exceeds revenue on most or all days the entire PPC campaign needs to be restructured.
Strategy 5.0 (Conclusion) — Need to Tend the Garden
Unfortunately you can just set the parameters and walk away from the PPC campaign. It requires constant attention. Things that work today may suddenly quit working tomorrow. Also there is always room for improvements no matter how well the campaign is performing.