Bid Management: Rule-Based vs. Portfolio

If you are working on sizeable search marketing campaigns you are probably well aware that managing bids for thousands of keywords manually becomes impossible to do efficiently. You need rules and algorithms to adjust bids for you and act and react to changes in the marketplace according to your predefined strategy. Adobe even thinks that “it is the models and algorithms that distinguish the players in this space.”

Adobe offers a variety of resources for search marketers in their online search centre, and among them a comprehensive white paper on portfolio-based bid management algorithms. Portfolio optimization stands in contrast to rule-based bid management. A rule-based keyword bid system would look at every single keyword mostly siloed from the rest of the campaign. For instance, you could sell DVDs online, and you could set the maximum amount you want to pay for a conversion to $20. Your system would take care of not stretching this limit for any of the keywords you are using. If the search term “harry potter dvd” would convert at around $21, your system would simply decrease the bid on this keyword or pause it. There is nothing wrong with that and this method would assure that “harry potter dvd” is converting at an acceptable cost. However, the portfolio approach says that this is thinking too short. Your overall performance might actually be better if you take advantage of the maximum amount of “harry potter” traffic, even if this means paying more for these conversions than you are willing to, and in turn making up for this with other keywords in your account that are converting more cheaply, and thus at the end of the day arriving at a higher ROI and conversion numbers than with a rule-based strategy. That means that a portfolio approach will look at all possible combinations of bids across all the keywords in your account, and then choose the best combination to maximize the overall outcome. It is clear that an extremely sophisticated and fast algorithm is needed to continuously adjust bids across a large account, as there are countless variables from position to CPC and CTR to be considered and all keywords need to be regarded as part of an ecosystem.

The portfolio approach can theoretically work very well – if not best – for campaigns with clearly defined goals and target margins and a predetermined schedule that is not constantly being changed or switched around. Note that there is also a cost involved whenever you connect to the Google AdWords API.

If you would like to read more about portfolio-based bid management, we highly recommend you give to Adobe white paper linked here a read.

The evolution of keyword length in search marketing

Posted by Matthias Wobrock on 28 May 2009

The evolution of keywords has progressed as users have become more specific in their requirements.

Our societies are rapidly being broken down into small areas of experts. People work in more specialised groups and more specific fields than ever before. The web supports this development, streamlining workflows and establishing world-wide connections to very specific information or niche areas of interest.

This trend can easily be applied to the commercial sector. If you view the web as a global marketplace, niche products also dispose of relevant sales potentials. Best practice example: Amazon. It achieves the majority of its profits through rare, long search term combinations. In the beginning of Search Engine Marketing (SEM), competition was relatively small. Today thousands of vendors offer essentially the same books, shoes or laptops and companies struggle to position themselves in terms of cost and creativity.

Expanded ranges make refined searches necessary in order to find the desired product. Additionally, internet users have become more skilled in using search engines and use more accurate terms to arrive at their destination faster. This development could even be partially owed to the relatively low precision of search engines. If “eBay Smartphone” does not bring the desired result, a search for “iphone 3g eBay app download” could do the trick.

Statistics not only show that the length of search terms is increasing (see bottom graph) but also that the percentage of searches with only one or two words decreases in relation to the total amount of searches. Searches consisting of 7 or 8 words are constantly increasing.

Interestingly, the number of total searches rises faster than the number of displayed paid ads, which means that the ad coverage on user’s searches is becoming a lot scarcer. Easy money if you manage to cover those complex search terms with your campaign.

From the first search for basic information to a more concrete purchase intention, search engine users may switch back and forth between generic expressions and long tail search terms. If you lose them on their course, you cannot influence their buying decision any longer. And if they use long tail keywords, they generally already have a more specific idea of the product they are looking to buy.

Bidding on long tail terms therefore offers two decisive advantages: on one hand, conversions are more likely due to the fact that the potential customer has already thrown a glance at a unique product. On the other hand, they are much easier and cheaper to rank for than general terms because the competition on that particular expression is a lot less: a better conversion rate for a lower cost.

However, these numerous keyword variants do not only need to be generated, but also managed, and their cost-effectiveness has to be proven. At Traffic Brand, for example, we use a custom-made system to identify and efficiently manage promising keywords to create additional value for our clients. We are prepared for this new trend and very curious to see how search engine usage will continue to change.