Products
The Challenge
Leads weren’t purchasing from the self-service site, so the lead quality needed to improve in order for it to be worth the sales team’s time to follow up with leads.
The trendline of the percentage of leads actually turning into purchases decreased proportionately to and corresponding with two main changes we made to the campaigns:
- increases in spend on campaigns on max conversion bidding, and
- increases in the number of conversion actions (steps of the purchase funnel).
The increased budget allowed max conversion bidding to bid too aggressively and drove down lead volume. The increased number of conversions worked initially, but with the increased budget, it started to confuse the algorithm and remove its focus on the most important conversion actions. We also noticed that purchases were mostly coming from locations immediately surrounding the processing centers the client had.
The Solution
So, we reduced the number of conversion signals for the algorithm to just two: topline leads (to maintain lead volume) and a down-funnel conversion that happens typically within a couple of days from the lead being received. This caused the algorithm to focus on a volume signal AND a quality signal that was highly correlated with purchase intent.
We also switched the campaigns to target CPA bidding. That allowed the algorithm to scale spend up and down while maintaining a relatively stable cost per lead. It also allowed us to eliminate dimensions (device, ad schedule, location) that generated lower lead quality.
A third part of the solution was to adjust bids based on top-converting segments. This included limiting geotargeting to only larger metro areas immediately surrounding the client’s largest processing centers, reducing bids on mobile, and increasing bids on high-converting timeframes.
The final part of the solution implemented over the course of the quarter was to separate higher- and lower-converting geo segments into their own campaigns to control budget to those locations and to have an additional lever to pull up or down to increase quality lead volume if performance from core geos started to trend below the monthly targets.
The Results
- 34% increase in new customer acquisition (amounted to about $77,016 in additional ARR)
- Cost per acquired customer decreased from $546 to $451 (17% decrease) – while scaling
spend up 10%
“The combination of target CPA, tighter geotargeting, and a reduced set of conversion signals provided several benefits. In addition to increasing our lead rate, it also helped us identify the causes of lead volume volatility that we had previously observed in our campaigns. Overall, the new strategy gave us more precise control over our lead acquisition rate, helping us to hit our monthly sales targets.” —Shiron Bell