Bidding technologies are not all alike, and the differences matter. While it’s true, many second-tier agencies use the same third-party software, the top-tier agencies all use their own. They have to; you can’t be the best if you rely on someone else’s toolkit, particularly given how central bidding is to success in paid search.

This is part 4 in a series on the required attributes/capabilities of a PPC bidding platform.

In the past, being able to bid down to the Keyword level was impressive, but now-a-days it’s no longer even sufficient. The most powerful systems need to go beneath the keyword level to the ad level for a number of reasons:

• Match types: Avoiding broad or phrase match totally is almost always unwise. People search using strange combinations of words, with odd permutations in word order. Trying to capture every possible permutation is impossible, even for us and we’re totally OCD when it comes to keyword construction. Given that and the fact that exact match traffic converts at much higher rates than broad matched, it’s important to have keywords running on both broad and exact, with the exact match keywords bid more aggressively. This is impossible with most bid platforms.
• Syndication Networks: Similar to match types, we know that traffic converts better on Google.com than it does on Google’s syndication partner sites (AOL, Alta Vista, Ask, etc). For our clients, we’ve found material benefit in splitting out Google.com campaigns from general campaigns, applying bids to each version of the keywords based on performance. Keyword level bidding systems can’t do this because they aggregate the keyword data.
• Geo-Targeting: Particularly important for retail chains with stronger presence in some regions than others, running the same keyword in different geo-targeted campaigns can allow these regional conversion differences to make you money, but only if your bidding system recognizes those differences. Financial institutions find this tremendously important in lead valuation, knowing that a “mortgage loan” inquiry from some zip codes are significantly more valuable from those in others.
• Copy Tests done correctly: Some folks evaluate copy tests simply based on Click-Through Rates. That’s fine sometimes, but you lose sight of the fact that the version that attracts more attention and clicks often sees lower conversion rates obliterating the QS and CTR advantages; Sometimes promotional copy during a promotion actually does harm! You can only see this if your tracking goes beneath the keyword-level. The same is true with landing page split tests.

Indeed, the truly powerful systems allow you to choose. There are times when it makes sense to force bids to be the same across some keyword versions — landing page tests, for example, you don’t want the early returns to sabotage the test before the data matures — there are other times when letting them float based on data is the most profitable. In any case, the days of keyword-level bidding as the gold standard are long-gone.

George

Bid System Imperative #1: Strong Statistical Foundation
Bid System Imperative #2: Deep Flexible Tracking
Bidding Requirement #3: The Feedback Loop
Bid Optimization Requirement #6: Human-Machine Balance