a fire hydrant is stationed next to a green wall

It’s unfortunate that funnels turned out to be the household item we use to describe completing steps online. 

The solution is not adding more money to the top of the funnel. The solution is widening the neck and making the funnel more efficient.

How do you widen the neck of a funnel?  Here are five tactics—rooted in optimization—to turn your funnel into a pipe.

Click to open in new window.

Click to open in new window.

1. Optimize keyword bidding.

Keyword bid strategy os the first opportunity to optimize the paid search funnel. By experimenting with different keyword bid strategies, you can influence the quality and conversion rate of your ads.

Let’s talk about a hypothetical business: a company that provides information for aspiring homeowners and renters in the San Francisco Bay Area. They are bidding on keywords that an aspiring renter would use in a search query. The keyword phrase to work with is, “rent apartment san francisco.”

To optimize keyword bidding, ask yourself this question: what effect does the keyword bid have on my key performance metrics?

  • Is this a suitable keyword?
  • Is the bid amount sufficient or too high?
  • What negative keywords go best with this keyword?

The goal is to answer these questions in a way that generates the most ad impressions, gives you the best ad position on the search results page, and maximizes the click through rate on your ad.

Decide what keywords to bid on by putting yourself in the shoes of someone with a challenge that your business can solve. Sometimes, users who visit your website through a search ad are looking to make a purchase immediately, but some audiences are looking to learn about your brand.

Test ideas for keyword bids:

  • Experiment with bid amount for different keywords.
  • Experiment with the effect that the type of bid, CPA, CPC or CPM has on performance metrics.
  • Test the long-tail terms you bid on. Long-tail terms are more specific keywords that have a lower search volume. For example, “2 bedroom apartment san francisco” is a long-tail version of the keyword “apartment san francisco.”
  • Test which negative keywords you add to your account. Negative keywords are words you specify that prevent your ad from serving. By adding negative keywords to your campaign or ad-group, you’re telling the search engine, “do not associate my ad with this search.”

In our example of a real estate company bidding on “rent apartment san francisco”, a smart negative keyword to add might be “average” or “historical”. Someone searching “average rent san francisco apartment” could be doing research.

2. Optimize the ad group structure.

An ad group is the container to organize your keywords and keep your account structures. Each ad group contains keywords, text ads and landing pages. The way you build an ad group determines the keywords your ads will appear in results for, the specific ad that appears, and the page someone will go from clicking your ad.

The goal in experimenting with ad group structure is to find the combination of keyword, text and landing page that converts more and costs less. Relevancy is key to a high performing ad group. The more congruent and related the messaging in your ad group are to the keywords people use to find your site, the higher the click through rate will be.

Remember, a high Quality Score is Google’s way of saying that your PPC ad meets your potential customers’ needs. The better you are at meeting the prospect’s needs, the less Google will charge you for the ad click.

To optimize ad groups, ask yourself this question: which combination of keyword, text ad, and landing page are best to lead to the highest CTR?

  • Keyword type—Should an ad group have only exact match, broad patch, phrase match? Or a combination?
  • Keyword included—What effect does including a specific keyword in an ad group have on conversion?
  • Ads in ad groups—What effect do varying text ads have on key performance metrics?
  • Segment by geographic region—What effect does showing different ads to people from .com versus .de have on key performance metrics?
  • Segment by device—Which ads should you pair with keywords for tablet users or mobile users?
  • Segment by time of day—Which ads convert best at different times of the day?

3. Optimize the ads themselves.

As much as you trust your gut, it isn’t always accurate. Experimentation will give you data to help find the highest performing ads. Click-through rate is highly influenced by ad text so it’s a great idea to be consistently split testing to find copy that performs best for your goals. Increasing click-through rates on your ads will benefit your paid marketing campaigns on two ways:

  1. Higher click-through rates will result in more traffic to your website.
  2. Higher click-through rates can result in lower CPCs.

Here are test ideas for each component of a paid search advertisement:


Headline: The first line of the ad and one that visitors usually notice first. (25 Characters)

  • Test Capitalizing The First Letter Of Each Word In The Headline..
  • Try making the headline a statement, a question or a clear marketing offer.
  • Does the headline include the keyword?

Description: Two lines of text that give detail about the advertisement, Here, advertisers commonly add a call to action to. (35 Characters)

  • You have 70 characters of text to experiment with here.
  • Vary your call to action.

Display URL: The line of text that gives visitors an idea of where they’ll go if they click on the ad. The display URL is not the destination URL (the precise location that someone who clicks the ad is directed to). (35 Characters)

  • Should you use dashes or underscores in your display URLs?
  • Where does a click take you?
  • What can a person infer from the link? If the link is ambiguous, will that increase key metrics?

Sitelinks: Sitelinks are a set of shortcut links at the bottom of the ad. (Max 6 links)

  • What effect does including or not including site extension links have on key metrics?
  • Which pages should you link to?

Note: Your ads will not always show sitelinks. It’s based on a trade off algorithm that Google calculates.

A lot of search engine marketers think their job ends with a click on the ad. The click is the last interaction that takes place in the search engine, but these marketers are missing the end goal, the full picture and goal of search engine marketing: acquiring customers. One last pivotal step remains in the customer journey that will ultimately decide if the paid search campaign was successful: the on-site conversion.

All the work and money you put into optimizing your ads and keyword bidding is for naught if you can’t turn that traffic into revenue.

funnel with cro

4. Optimize the pages you send paid traffic to.

“On average, 96% of the time, people click on these ads and do not perform the action that marketers desired them to do,” says Wordstream co-founder and CTO, Larry Kim. “They visit the site and don’t convert to a lead or a sale.”

A lot of people click ads, but the vast majority of that traffic doesn’t convert. This seems counterintuitive. If click through rate is high, that means traffic to landing pages is high. Shouldn’t conversions on landing pages be high too?

Not necessarily.

Most landing pages are poorly designed to convert paid traffic, or they’re non-existent. Many marketers send paid traffic to the homepage or pages on the website that lack opportunities to take action.

What makes landing pages a no-brainer for A/B testing? Landing pages may not be part of your regular website. They are URLs that exist for campaigns. They are keyword-focused, topic-focused, designed for taking an action. So unless the “About Us” page has a signup form or checkout button on it, it’s very unlikely you would ever use it as a PPC landing page.

You will use similar landing pages for a number of your search terms. Because you can have a lot of them, they can all look almost the same. So optimize a couple and apply those lessons across the others. A gain on one landing page could exponentially increase ROI across your entire paid search spend. Of course, it’s not guaranteed that what works on one page will work on another—that’s why you have to test. Check out a related article I wrote on this topic, Why You’re Crazy to Spend on SEM but Not A/B Testing.

5. Convince your team it’s worth it.

The final step to thinking end-to-end about optimizing the SEM funnel is getting buy in from stakeholders on your team.
The Roadmap to Optimization

Download our in-depth guide on building a testing team at your company.

If you’re a team of one, then you have a great place to start. In other scenarios, you could have a couple people on your team who handle different areas of paid advertising. One person might be in charge of SEM, another person manages content on the website and a different person controls the budget. If that’s the case, then we have advice for you: Start a task force and get people involved.
Launch the task force with a stakeholder meeting. Invite everyone who could be involved in launching A/B tests or analyzing data. Get people excited by trumping their intuition. Show them a series of A/B tests and prove that gut instincts are not always right.
“One of the most interesting parts about CRO is the competitive aspect,” says Eric Fisher, senior manager on the Conversion Optimization team at Performics. “Promote this aspect of CRO to get people involved and make the connections within your company. Start growing that presence and visibility into testing because in order to have an effective testing program, you really have to have everybody on board.”

In another meeting highlight the connection between SEM and CRO with numbers. Look at two numbers, the amount of money you spent on paid search marketing last month and the conversion rate across all of your landing pages and revenue you’ve generated from that traffic directly.

Sit in a room with your AdWords account on a screen and benchmark your numbers. How are you tracking on key performance metrics? Identify the lowest conversion rate and come up with five hypotheses to fix it.

If you don’t have a software that allows you to do A/B testing today, then consider shopping for one and make other people part of the demo and decision-making process.