Growth Mode: Lead Generation
When running Lead Generation campaigns, we take a page from our DTC practice and focus, as much as is possible in a post-iOS14 world, on post-lead events. A lead has no inherent value, just like a site visit has no inherent value. The value is dependent on our ability to convert that lead into a monetized customer.
Audience and Creative
The first step in the conversion process is audience selection and creative. We think along a spectrum, from LEARN to EARN. In the initial phases, we’re running fairly granular audiences against each other, to learn the relative performance differences between them. Based on performance, we’ll scale, kill or merge accordingly as we shift the focus to the EARN end of the spectrum, in which we focus on setting campaigns for long-term scaling and algorithm signaling.
We take a similar approach to Creative iteration. For creative, we try to test very large differences against each other, before honing in on the winning variants and iterating within that. Imagine another spectrum, from A to Z. Rather than test between A/B/C variants at this stage, we try to test between A, M and Z. We don’t want to test very similar ads, with differences only in font or button colors. Instead, we want to test very distinct creative concepts, identify a winner and then iterate on minor changes. So in the example above, if we found that the “M” creative concept performed best, we’d develop L, N and O variants to try to improve that Champion variant.
Our goal isn’t to find the “best ad”, it’s to find the best ad for each audience. We want to see and react to the peaks and valleys, not the plains in between.
Next, we focus on the on-site experience, primarily through the Landing Page and the Lead Form. For Landing Pages, we take a similar approach, developing distinct creative and experiential approaches before honing in on the winner and optimizing within that creative pillar.
For creative and landing pages, our goal is broader audiences and wider testing apertures. The Lead Form is an exception to that process.With Lead Forms, we take a much more focused and prescriptive approach, one that is highly dependent on our client’s sales process.
Typically, we evaluate the funnel based on its ability to move users down the funnel, as measured by conversion rate. A short form with few required fields will have the highest conversion rate… but conversely, it maximizes the number of low quality leads (lookie-loos) who make it through to the next stage. If your next stage is a self-serve demo, that’s not much of a concern. But if the next step is a live demo, you may want the form to separate the wheat from the chaff. In other words, we’re so used to sanding down points of friction along the lead journey, but sometimes friction can be a useful filter.
The form structure can balance between the benefits of a very open, non-restrictive form (lower CPLs, higher conversion rates) with the degradation in lead quality. If you removed fields, or made all fields optional, the CPL would likely drop, but so would the quality.
If you’re trying to minimize the waste (in this case, we define waste as the form completer who doesn’t get on a demo call), then add in more fields to make it a higher-commitment process that aligns with the higher intent of the leads who do schedule a call.
Typically, you expect to see a decrease in the Form conversion rate, which leads to higher CPLs. But among those who do put the extra work in to complete the form, a higher percentage will be higher-quality (ie., will schedule a call). On-Facebook Lead ads are the perfect example of low-friction/low-quality lead gen.
Our general recommendation is to not be very restrictive at all – to do everything in your power to maximize the number of leads generated at the lowest possible cost and then engage automation to nurture those leads down the funnel to a call, followed by post-call automation to maximize the number of call participants who close. Automated emails, self-service call scheduling, email follow ups, and paid retargeting (suppressing call participants) all work together to take those at the top of the funnel and matriculate them down-funnel. That recommendation is subject to your particular internal capabilities and SDR bandwidth, of course, so it makes sense to put some constraints in to control the volume. But in terms of long-term plans to build out the funnel and leveraging all your marketing channels – internal, organic, free, earned and paid – maximizing the amount of first-party data under your umbrella does that best.
Typical Lead Gen agencies focus on delivering leads under the target CPL, and then walk away, leaving it to the client to worry about converting the leads to Closed Won. We know that a lead holds no inherent value until it converts, in the same way a visit holds no value until it converts to a purchase, even if that conversion happens days or weeks later. Many times, analysis will look at CPL or CPC and use that as the basis for future CPLs – but then they end up optimizing to lower quality leads in the case for the lowest cost. CPL is fine as long as we keep the end goal in mind: CPCW.
We work with a SaaS client who wanted to ramp up their lead gen volume signficinactly. When modeling their goals out, they looked at historical (primarily organic) leads, and directly extrapolated the value and cost metrics out within a paid campaign. They looked at their historic leads and divided by the value of those leads, but such forensic valuations don’t factor in the original context – they’re using organically-generated leads to model paid efforts, optimized to a CPL. Those who have run on-Facebook lead ads know that the value of those leads is not the same as organic (or even paid leads that are generated onsite) and so there’s a danger in over-extrapolating.
If that analysis then shows that the value generated by a lead is $800, that doesn’t mean you should start CPL-optimized campaigns, using $800 as the benchmark value, because you end up getting a lot of sub-target leads that don’t convert as well as the more organic historically-sourced leads. With this client, we created a Performance Framework with specific CPL, LTV and ROAS targets based on historical data, but adjusted for the change in marketing mix and organic:paid split. We found, through A/B testing, that optimizing to CPL generated a much different conversion rate and pLTV than optimizing to ROAS or Traffic, and each of those were different from organic sources.
Ultimately, Lead Gen is an interim or proxy goal along the way to the true goal – Sales Gen.