This post originally appeared as my column on Search Engine Watch. The idea of a good paid search account is a myth. An account is simply an aggregation of campaigns. Campaigns combine ad groups. Ad groups roll up text ads and keywords. Keywords and text ads exist only to attract search queries. Effective PPC analysis is about finding the segments in your account that deviate from the norm. In many cases, you’re leaving opportunity on the table. The “norm” is all about context. Winners and losers are different than their counterparts. Here are 6 useful tips for setting context and hunting down losers:
- Internal Benchmarks
- Against Potentials
- What’s Changed?
- Trend Charts
- Against Goals
- External Benchmarks
1. Internal Benchmarks
We’re all obsessed with comparing ourselves to other people. How does my conversion rate stack up to my competitors? What’s their clickthrough rate? There’s value in competitive benchmarks (see number 6), but getting competitive data can be expensive, time consuming and not always as actionable as focusing internally. Setting internal benchmarks, i.e. judging you vs. yourself, is the most realistic way to assess your potential. Take text ad testing as an example. Text ads play a central role in determining your clickthrough rate and attracting an audience that will convert. Judging a text ad is all about comparing its performance against peers in the same ad group or campaign. You’re trying to find the combination of offer, headline and other factors that drive higher CTR and conversion rate. You need to monitor and ostracize your underperforming text ads.
Benchmarking against an average can also help you spot weak segments. Google Analytics has a useful visualization to show how segments like keywords compare to the mean for various metrics. Here’s an example of charge showing bounce rate by keyword vs. the site average from my blog (in this case, lower is better).
2. Against Potential
In some cases, you can spot losers by comparing their performance against a known threshold. For example, Quality Score has a defined range:
- 1-4 is poor
- 5-7 is okay
- 8-10 is “great”
In plotting out the distribution of Quality Score for a number of clients, we’ve found that 7 is the average. Anything above that is a reward, anything below that is a penalty.
Armed with that information, you can monitor your account for keywords that slip below the 7 mark. In other cases, there isn’t a defined target, but there are best practices (or good practices) you can pick up from trial and error or the feedback of people with larger budgets, experience or sample sizes. For example, you could choose to focus on building out negatives for keywords using broad match that have an excessive number of search queries. Or, you could prioritize analysis on campaigns where Impression Share Exact Match is low. If you don’t have a system to alert you to segments that might be missing their potential or violations of best practices, a simple spreadsheet with conditional formatting can help you quickly spot the losers.
3. What’s Changed?
As Search Engine Strategies New York keynote speaker Avinash Kaushik said:
“While we obsess about our brand terms and our top ten key phrases the reality is that the long tail of search means that our organic and search campaigns focus on tens of thousands or hundreds of thousands of keywords.
One effective strategy to deal with this purely data problem is to focus on what’s changed.”
Focusing on what’s changed makes it easier to manage large, high spend accounts. Instead of having to pore over reports to spot the losers (or winners), you can focus on segments that are spiking up or dropping rapidly. There are three parts to building an effective what’s changed report: 1. Segment – Campaign, ad group, keyword, text ad, etc. 2. Metric – Pick the data points whose change you care about 3. Time Period – You two time periods to define a change. I suggest you start with Ad Groups and Keywords. Changes at the campaign level are more likely to send you digging deeper anyway. Keep your metrics as actionable as possible. Focus on KPIs like Gross Profit, ROI or Cost-per-Acquisition. Week over week and month over month comparisons are a good place to start. You’ll end up with a report like this:
What’s Changed Report built with ClickEquations Analyst
4. Trends Charts
Sometimes, just trending your data over time makes it easy to spot issues. In this chart, I used dual axes to show leads over time (left axis) and the cost-per-lead (right axis). You can quickly spot issues on 3/2 when Google leads tanked and 4/6 when Yahoo leads decreased. Now you know where to drill down.
Every great dashboard has some version of this chart. Just like you segment the data, you can take it even further by segmenting your charts.
Sparklines let you embed mini-charts in your reports. It’s a great way to see the trends underneath the data when you’re presenting tables. Here’s a sample chart that shows the absolute value, week/week changes and 12 week trends for a variety of metrics for Yahoo. This chart was built with Microcharts. Hat tip to my coworker Matt for his Excel wizardry.
5. Against Goals
One of the most insightful ways to add context is to contrast your progress vs. your own goals (assuming they’re realistic, of course). A pacing report is an easy way to get started. (Thanks to Josh Dreller for the name). First, set your goals. Each time period (say, a week), you report on that week’s performance against those KPI’s. Then, you add the context of results-to-date vs. those goals. Next, and most important, you project the likely performance for the end of your time period (say, a month). Start with the simplest projection. Take the average daily results for each KPI (ex: 10 leads per day, $200 spend per day) and multiply it by the number of remaining days in the month. Here’s an example: By their nature, pacing reports work well at the summary level. It’s much less common for PPC managers to have segmented targets. If you have enough account history, however, you might be able to set goals at the campaign level. In that case you might consider using a bullet graph in your analysis. You can easily scan and find underperformers with the additional context. Source: Charley Kyd at ExcelUser.com.
6. External Benchmarks
It’s nearly impossible to get conversion rates on your competitors’ sites, but there are a number of useful tools that can help you get intelligence about which keywords your competitors are buying, CPCs, etc. These tools typically work by using panel based research or by scraping search engine results pages. Compete Search Analytics Compete.com (now part of TNS) offers an accessible Hitwise alternative for seeing two key data points:
- Who’s winning traffic for keywords and
- Which keywords drive the most traffic to your competitors
They have a free version of Compete Search Analytics that you can try without signing up. [Disclosure: Compete gave me a free Pro account for evaluation purposes.] Here’s a sample report for the word “laptop”
The Search Monitor The Search Monitor has a lot of useful features for competitive intelligence and, in particular, affiliate monitoring. They also have a unique report that lets you see what dayparting strategies your competitors are using.
AdGooRoo AdGooRoo has a suite of PPC and non-PPC tools. Agencies in particular will like their spend benchmarks, which help you see how much your competitors are budgeting for PPC, so you know whom you’re competing against most often.
Price points vary, though most tools offer reduced feature free option or some full featured free trial for a limited period of time.
No Losers Allowed!
You don’t need to use all of these techniques to be successful at PPC. But, the larger your account, the more competitive your vertical and the higher your paid search spend the more likely there are losers who are dragging down your performance. It’s time to kick them off the cool kids table.