We’ve all been here before….
Sales: “We need more leads!”
Marketing: “Why don’t you call the ones you have?”
Both will fight to the grave that they are right in their stance.
So who is right?
For a long time, I always thought the answer was that the marketer was right. Shocking, since I’m a marketer driving the best leads on the planet, obviously…
But as time progressed and I spent more and more time bridging the gap between marketing teams and sales teams, it became apparent that we were both right. It was clear to me that there were, in fact, good leads and bad leads.
But what wasn’t clear to me was how to give them only the cream of the crop.
There wasn’t one magical channel that held all the value, and if I shut something down, there was always a case of throwing a baby or two or ten out with the bathwater as all good impressions drive value to the end goal of success.
Every business I have seen the guts of has run into this problem, and every company has attempted different remedies depending on the stage and level of sophistication of the business themselves.
The information here is intended to be a valuable hack for any marketing or sales professional to help put together the initial set of information need to start cutting down on dials, and driving higher ROI immediately.
Taken to the next level and paired with a high-powered data team, this framework is what led to driving 100% of new business with 40% fewer dials at LendingHome.
But, before we get into the “how”, let’s start with the “why”.
If any of these sentiments make you nod, sit up in your chair, or make you shake your head, you’ll definitely want to keep reading through.
Let’s get into it.
The Vicious Cycle
Here’s the most common and vicious cycle that I’ve seen play out at different organizations when it comes to driving growth. With enough time spent around SaaS startups, just about everyone has seen some form of this “leads vs headcount” dynamic play through.
SDR teams always want more leads, but there’s never a great idea of what the “right amount is”. If budgets are increased, an SDR team is swamped with leads of varying quality there is undoubtedly a point of diminishing returns.
They may be technically producing “more”, but it’s not in proportion to the incremental marketing expenses that are coming along with it.
Here’s why we tend to see that unfold:
- Without scoring, SDRs spend an equal amount of time on each lead or apply their own scoring which causes them to not call certain leads at all.
- As paid budgets increase, the overall quality of lead volume tends to diminish (let’s be real marketers, it is incredibly true!)
- As total leads increase, dials per lead diminish.
- As dials per lead diminish, conversion rates to qualified customers diminish.
- As rates to qualified customers diminish, ROI diminishes with it.
- As ROI diminishes, everyone is under the gun.
At this moment of time, there are usually a couple of options.
Do we turn off spend? Hire more SDRs? Are we happy with this ROI and continue on our current course so we can drive extra business?
When possible, it tends to be advantageous to invest more in acquisition over hiring additional fixed headcount cost to take on work if they can accomplish the same objective.
This keeps the balance sheet a bit more flexible. It’s much easier to adjust budgets up and down rather than headcount.
Below is how to spend more, keep headcount flat, and make sure you’re extracting the right ROI from your leads.
Know What to Solve for
At the end of the day, in any B2B and start up optimization, the true battle is with time.
As humans, we’re limited to a finite number of actions in any given day. Figuring out how to address this problem and make customer facing teams “super humans” is the true challenge.
The first time I went through this exercise, I thought what I was attempting to do was pin point the exact perfect lead.
But, as I dug into the data, I kept getting stuck.
There were a handful of leads that I could perfectly predict, but if I threw everything else out I’d need to ax the sales department by a large amount since I’d never be able to feed enough mouths. Not to mention my cost per MQL would be through the roof…
What eventually became the key for me was flipping the model on its head entirely. Instead of trying to predict the perfect lead, I found that there was far more value in predicting the worst leads. This unlocked a much simpler path to driving value.
Getting the Information Needed to Score
Structuring data in a way to make actionable decisions is the core of all things performance marketing. The core ones used for this purpose are:
URL structure – This is the heartbeat of all paid performance campaigns. Applying attribution throughout the customer journey is key. If you’re new to the game and need a good reference of where to get started with this, here’s a good start.
Funnel milestones – in B2B, conversion cycles are long and not nearly enough folks make their way to the end of the line. Make sure you pick a conversion point further up in the sales cycle that is indicative of value, but not so sparse that there isn’t enough meaningful data to run a regression against. For the sake of this exercise, I’ll be using a sample of leads that were “converted” by the sales team.
Form questions – every question in every form matters. Make sure they do, and that the data you pull from them is in a location for you to analyze and evaluate. Don’t just ask of information from potential customers for the sake of asking!
Additional fields to score against – The list here goes on and on….total visits, interaction with content, email domains, etc. For the example I’m going through here I don’t use any of these features, but in the graduated models where scores are live calculating I definitely am!
Running the Regression
As discussed previously, let’s remember the core objective here:
- Remove as many bad leads as possible…
- While removing the least amount of conversions as possible.
At this point in time, Excel becomes a marketing or sales professionals best friend. The easiest way to get through this is with pivot tables. If a brush up is needed, here you go.
Step 1: Clean the Data
Every data set is funky in its own right. Conversions could be counted in different columns depending on what is counted as success: ie, a “demo set” vs “converted” might both be great signals of funnel conversion, but they’re counted in different columns or names when exported. I like to add a simple column that is simply “Success” with a single value of “Yes” when true. This allows for an easy way to count the numerator when trying to count conversions.
Step 2: Pivot all the inputs to look for outliers
At this point in time, it’s time to start pivoting everything you have to start to measuring success. The calculation here is pretty simple. Success / Count of the Population for that parameter.
For every unique feature identified, see how many conversions came from each unique response value.
The hope at this point is that there are some features in the data set that are emerging as standouts. In the instance below, when someone answers “Within 1 Month” to the “Timeline” question, it converts at over 2x the next best option in the data set.
Step 3: Set some initial assumptions
Making a calculator to help go through different permutations allows for a quicker path to an ideal outcome.
Once the features that are going to be valuable have been identified, it’s helpful to build a calculator that lets you see how your scores are impacting different models.
Here are some outputs based on different weights to varying features, but feel free to copy them out of this sheet and play with on your own to see how different permutations impact the results.
Step 4: Tinker with the calculations
Now the question is, “how do I start testing scores”. Once again, there are some very scientific ways to get this done. Since this particular solution is a hack, we’re going to go with good ol’ fashioned trial and error. Going back to the main thesis of “remove as many of the bad while not eliminating the good”, I’m going to tinker until I find the right combo.
Now that there is a baseline in place, try to adjust the calculator to make an impact on the ratio of leads eliminated to conversions eliminated. This involves two different efforts. Since I am isolating lead score of “1” as my cutoff threshold, I am trying to 1) push as many conversions as possible from that score group into score group “2” and 2) Trying to pull as many leads back from score group “2” that do not have a conversion into score group “1”. Here are a couple scenarios that I tested out, showing instances where I both gave some features bumps, and some I applied negative scores to.
Step 5: Pick a winner
For the scenarios listed above, the bottom one is what I found most attractive.
While it is sacrificing 2% of conversions, it is eliminating 19% of the total leads that need to be worked. On this data set, that equated to 15% of the total dials that the SDR had to make to drive only 2% of total conversions.
Building Out the Functionality
Once there is a model that’s looking good, it’s time to put it to action.
Take the scores from the features that have been identified and build them into a workflow sequence in a CRM.
Write this score to a new column into SFDC or equivalent solution and now there is a score column to create views from.
This should be a layup for a sales or marketing ops expert in terms of implementation. Bonus points if they re-score the model every night (or more frequently, depending on how quickly actions pile up with user behavior).
This way leads won’t be permanently painted outside of the scoring threshold and it gives automated nurture streams a chance to push customers over the edge into a workable contact.
Selling it to the Sales Team
This is where it gets tricky.
When it comes to leads, sometimes subjectivity and “I feel” logic seems to defy math. The sentiment is always that more is more. But, as we know, 100 dials spread across 50 leads goes a lot further than 100 dials spread across 100 leads. Here’s how to sell it:
You didn’t invent scoring – Every rep I’ve ever met interprets every lead differently the first time they open it. This is already a tactic that they are doing in their heads, so they know that the logic is already sound. All this process does is show them that they were right and that there’s a next level of ability to apply their logic.
Include them in the Process – Rather than blindly look at data to identify where to start, try to get a bit of a head start. Sit a couple of the top reps down and open a few leads with them. Try to see what they look for and identify trends to take deeper dives into the data set.
Lift Open the Hood – walk the sales team through how the conclusion was landed at it. Show them how deliberate the process was. They should make connections based on what they already deemed as “great” leads on their own.
Drive home what matters – Making Money & Growing the Business – Lead scoring isn’t all about removing leads from pipelines. It’s about making sure the leads that need to get worked are receiving the proper attention. A great outcome of this approach is to eliminate the bad leads in the pipeline and then spend a little bit more to replace them with good ones. This is good from them, it’s good for you.
More Ways to Apply the Score
Once you have the score itself in place, there are several different ways to apply it to enhance a business.
Sales Automation – Just because a lead doesn’t pass the initial score threshold right off the bat doesn’t mean they are ignored forever. Once again, it simply means that it’s not worth the time dedication without the customer showing a couple more signs of life. Building a nurture stream, either from the company itself or from a sales rep, allows those that aren’t getting called to have a positive experience with the brand and give them the appropriate steps to move down the funnel on their own.
An Alternative to Altering Budgets – No longer is their the constraint of sales headcount being a factor for whether you have to accelerate or withdraw budgets from campaigns. If an SDR gets promoted and leaves the staff thing, no worries. Just raise the threshold for the lead score, pass less through, and the remaining reps will still be able to hit their call SLA’s. In many cases, it allows for a much stronger argument to raise budgets overall since pipelines will be more effective.
Driving Further Segmentation in the Lead Set – What we have discussed so far is a very binary “pass” or “fail” mechanism. From there the next step is to drill down into the next level of segmentation for leads that have passed. Tactics here include different dial SLA’s, different product experiences, and different messaging in automation. I’d recommend slicing off a portion of the highest scores and treating them like must win clients.
Testing Alternative Sales Tactics – Many companies seem to have starts and stops when it comes to outbounding and how different “tests” are set up in this capacity. One of the ones I consistently see fail is trying to have reps both inbound and outbound at the same time. Use this extra gift of time and bandwidth to allow the room to carve out a rep to conduct a solo outbounding effort for a couple months to gain true efficacy of the channel.
List Reactivations – The new lead score model that has been developed doesn’t only have to work for net new leads moving forward. Once you run the regression, upload the score to all leads in the database. This will identify the top potential candidates sitting around that may not have been identified previously. Making more out of the dollars that have already been invested will always make your CFO happy!
Change the Questions you Ask (or the way you ask them) – if you have identified a couple of fields in the regression that stand out, try to adjust lead forms and questions to extract deeper segmentation to a particular question. For example, if you were previously asking a “Yes” or “No” question, what can be done to split that up further to dissect the intent of the customer and drive more meaningful segmentation for a sales team to act upon? If there aren’t any standout questions in your flow where distinct positive or negative behaviors are occurring, I’d encourage revisiting capture forms at a macro level.
Wrapping Up
Remember, the more leads that are filtered out, the higher the quality of the scores that remain. While it’s tempting to regress to old habits to “let more leads in”, remember that they aren’t worth it! It takes a disciplined approach to make this work without reverting back to old habits.
At the end of the day, in any B2B and startup optimization, the true battle is with time. As humans, we’re limited to a finite number of actions in any given day. Figuring out how to address this problem and make customer facing teams super humans is the true challenge.
Happy dialing!