These are the results that show at the end of this section. As will be demonstrated, they are impossible to realize or recreate without fully utilizing internal data.
One thing about niches is that if you aren't running an account with pre-existing high quality data, you don't have as many external resources to rely on. Long-tail keyword data within niche verticals is often scant and occasionally non-existent. You can use a range of tools to start gathering data, but it will take a long time to have enough reliable data to work with.
In this case, I had to use data from a previous period that provided unsatisfactory results for the client. Workarounds made this problem tertiary; but from beginning to end, research started with looking inward. Successful long-term strategies rely on utilizing internal data as much as (if not more than) external data.
Month 1 | ACCOUNT AUDIT | Audit provided immediate improvements and guided roadmaps for strategy options going forward
One thing about niches is that if you aren't running an account with pre-existing high quality data, you don't have as many external resources to rely on. Long-tail keyword data within niche verticals is often scant and occasionally non-existent. You can use a range of tools to start gathering data, but it will take a long time to have enough reliable data to work with.
In this case, I had to use data from a previous period that provided unsatisfactory results for the client. Workarounds made this problem tertiary; but from beginning to end, research started with looking inward. Successful long-term strategies rely on utilizing internal data as much as (if not more than) external data.
The audit made the need for certain strategic changes inarguable. Others, however, would take experimenting. The legal vertical is expensive, and when your competitors are spending 7-10x more than you, experimenting can come at the cost of results. Because this client had poor experiences with PPC in the past, it was safe to assume that any explanation for a month-long leads drought so early on would be of little use, no matter how valid. I needed to find money in the budget; enough that would reliably cover the cost of experimenting.
Objectives
Find enough money in the budget to experiment
Use the audit as a source of ideas
Prioritize the most pressing questions and generate the internal data to answer them
Do both of the above without causing radio silence from PPC leads
This client preferred phone call leads to contact/lead form submissions, and voicemail leads had a significantly lower conversion rate at their firm.
It was clear. The best place to save money was ad spend outside office hours. But first I needed to know what performance would look like in concrete numbers. What kind of search traffic did evening yield? Were late-night users lower intent leads? Were there differences between practice areas in this regard?
From the existing data across the three main practice areas (Social Security, VA disability, and Railroad Retirement) the time trends were indeed wildly different for each. Ultimately, this research allowed us to save 48.6% of the budget - nearly half - without causing any noticeable damage to existing performance levels.
Month 2 | OPTIMIZING BUDGETS | By cutting out low intent dates/times for each practice area, I saved half the budget for experimenting
Some of the strategy I implemented can be shared below. Namely, how external data was used to leverage greater outcomes for their firm.
TIME TRENDS
YoY trends provide critical data when building a long-term roadmap and budget. By skipping August and September - a decision made due to data from the previous three years - we were able to expand the budget for the October spike. This was a controversial decision at the time, but ended up paying off significantly for the 3m period.
October was a very good month.
SEARCH VOLUME
Search volume can be tricky when it comes to exact figures. The most reliable source of data would be Google if GKP gave specific numbers instead of ranges.
By pulling from trends across multiple data sources (SpyFu, SEMrush, Moz, etc.) and leveraging internal data - namely impression share metrics, lost impression share metrics, and data from auction insights - I'm better able to reinforce the validity of volume estimates.
Towards the end of our year of working together, I put together a comprehensive YoY comparison of the current year and the previous one. As mentioned earlier, one of the biggest problems was that the client wasn't seeing Google Ads conversions turning into real-world leads. A lot of this was resolved through crafting a much narrower, high-intent acquisition funnel, as well as major improvements to attribution.
However, since that tracking couldn't be applied retroactively, the best way of getting a read of the improvements to their account was by sifting through the search terms across both years.
I know what you're thinking: Auseel, that has to be an insane amount of search terms... And you're right. Somewhere around 50,000. Using my own personal time, I sifted through them, filtering out low intent queries; sometimes by rules and often by hand. Because the enclosed information is more usable than the rest of what is included in my portfolio, the details of my methodology will only be shared with serious inquirers who contact me personally.
After I was through applying the changes equally across the data from both years, the results showed progress that further satisfied an already happy client. Below are the summary results of the YoY comparison.
Month 12 | 2021 vs 2022 | A lot of champagne was had at a Frank Stitt restaurant the weekend after this report was shared, courtesy of the client
Importantly, in all this time, we properly trained automated bidding, feeding it accurate conversion data on an ongoing basis. We experimented with smart bidding at Month 1, Month 3, and Month 6. The first two attempts showed that the account was nowhere near ready, with the strategy bringing in junk traffic. At Month 6, the quality of traffic improved, but (1) it wasn't meeting the volume we were generating at manual bidding, and (2) our CPA was marginally higher.
As you can see, at Month 12, smart bidding was running both more effectively and efficiently.
There are a lot of strong opinions about automated bidding in the legal vertical out there, although fewer now than there were a few years ago. I'm of the opinion that it works for most practice areas so long as they have the volume of data and expertise to train the machine well.