Sourcing new outbound leads is a never-ending endeavor for sales and marketing. Prophesies of cold calling being dead have not come true if only for the simple reason that prospecting through all channels must be on the table to fuel the engine of sales growth.
Companies typically procure lead data from four main types of leads:
Internally-generatedleads: Referrals, word of mouth, events, etc.
Intelligence-based leads: Newsfeeds, industry alerts, personnel changes, etc. InsideView is one good example of this, but we believe LinkedIn also fits this mold (the evolution of this is exciting)
Special or vertical lists: Trade associations, commerce groups, organizations operating with a geographic charter
Compiled lists: The likes of D&B and InfoGroup, including credit files
But as any sales or marketing manager knows, simply dumping more records on a sales rep is a thing of the past. Lead nurturing and scoring are the norm, wherein prospects are nurtured until they raise their hands as hot leads, and are then forwarded to sales. In addition, predictive modeling that identifies the most likely prospects based on the “ideal customer” profile must be part of the mix.
It’s clear that the prospects at the top of the funnel – the ones you might pay for – are not sales ready. Further criteria and filters must be added which reduce the number of truly viable sales-worthy leads. Let’s take a look at how the leads get whittled down in the pipeline – we call this the “prospect waterfall:”
The net result is, if you are paying for leads and then “throwing away” 55% of the records (typical from our experience), your true list cost goes up by almost 125%. Add the additional processing cost, sales rep fatigue and opportunity costs, and you can see why your prospecting results and ROI are not what you had hoped for!
For best ROI, pay only for what you use.
If your current processes throw away prospect records that you paid for before the sales reps see them, it is time to reassess your lead generation program. Or as one of our clients put it, “stop buying leads until you figure it out.”
When you change your lead scoring and lead delivery using predictive analytics, don’t forget to train sales reps to think different as well.
It’s well-known that salespeople don’t qualify leads, they disqualify them. The more leads provided, the faster leads seem to get disqualified and bounced back in the holding queue. Reasons could be due to lack of data (such as invalid or no phone number), bias (“can’t possibly be a large enough deal”), or simply attitudes (“the more I close out, the sooner I will find something that works”).
This approach churns through leads, resulting in significant cost of acquisition and processing resources – both human and machine. If your reps’ closeout rate is 50%, your net cost is twice the initial cost! Marketers and sales leaders often respond by finding ways to deliver a greater number of qualified leads faster.
Predictive models are often used to score and deliver ideal prospects from a larger universe into outbound lead gen programs. Predictive models increase productivity and the ROI of achieving specific outcomes such as getting appointments or sales, moving newly-acquired customers into repeat customers, and improving cross and upsell.
But caution: predictive models may produce leads with characteristics that are counter-intuitive.
This is where you get hot leads to handle with care. By their nature, scored leads will be a lot fewer — typically 20-40% of available leads. High-scoring leads may have attributes considered not desirable, like the number of employees. These leads may also not have some fields that reps typically look for, like contact title. Yet these leads have been predicted to produce a high level of performance.
This leads to a few lessons:
Teach reps to look past some traditional criteria and look at the prospect holistically.
If reps continue their old practice, they may close out leads more likely to convert. And because these leads are fewer in number, upon exhausting them reps go on to scour other lead sources that perform significantly worse, as we’ve found upon back-testing. This socks a double whammy to your performance and productivity. For maximum performance, do not allow reps to pull leads from other sources unless all the scored leads have been acted upon.
Leads that score highly for one outcome, such as getting an appointment, are also more likely to produce a sale, at a higher average order amount, and with more add-ons. Train reps to nurture these leads to achieve such multiplicative sales value.
In short, predictive analytics leads you on a fundamental shift from tactical to strategic thinking. Predictive analytics may deliver fewer but better leads – be careful to not burn through them. Instead, profit from their sales potential.
Inside sales teams are at it every day, making thousands of calls to prospects, seeking an appointment or a sale. They are given scrubbed lists with contact names, job title, phone numbers and a good luck pat. On the back end, managers track calls, appointments and sales. The cycle continues when reps deplete their assigned leads and it starts over again.
There are times when this well-oiled prospecting machine can under-deliver – and you may not be aware. But there are easy fixes. Here we explore three cases, and discuss how to overcome these challenges.
TESTING NEW MARKETS: When looking at sales data, you may find strong traction among companies that don’t fit the best customer profile — or at least what the rep thinks is the best profile. This finding is usually uncovered by in-depth profiling, micro-segmentation or modeling analysis. These customers may not be among the largest customers, but you find that they purchased numerous units of products that fit specific needs.
As an example, say you are selling technology products, and religious institutions are not known as leading technology buyers. But recently we came across a church buying hundreds of iPads for one of its programs. This is opportunistic entry into a market if other religious institutions have similar programs. What you can do:
To aid sales calls, share historical product purchase data with reps so they lead with iPads and related iPad cross-sell products instead of laptops, printers or software.
Show reps the sales potential of these calls when they may understandably question this. Pull examples to show similar sales to this market and the value of the deals.
To help target this new vertical, it is important to match the highest product propensity to the leads in this vertical.
MISSING INFORMATION: Even when providing a scored list through predictive modeling, some information valued by sales reps for their calls may be missing.
In one case, sales reps for a technology provider perceived the presence of a website as a surrogate for computer purchases. In our B2B database of 14 million businesses, this field is only available for 22-25% of sales prospect records. In our scored file this info was available for 35% of records – certainly an improvement, but reps were concerned about the two-thirds of leads without this info. The scoring rated these leads as strong prospects but reps were uncertain about calling them. What do you do?
Have them do a quick Google search on company name and state. The company may have reviews, social media presence, etc. that would otherwise confirm they’re “tech-savvy,” and if there’s a website it will likely turn up.
Show cases where leads without this information have progressed through the pipeline and converted successfully.
Explain that the process of data collection, while well-defined, is not always perfect. Set up specific steps they can follow to cross-corroborate other fields and deduce the missing information.
PROSPECTS HEAVILY SKEWED: Recommendations from predictive modeling may favor market segments that reps do not normally associate with large sales. But the model shows easy smaller sales opportunities exist.
One of our projects produced recommendations where the prospect company size varied from 100 employees to as low as 8. Most records were towards the lower end, reflecting the business universe and perhaps implying employee size was not an important predictor of sales. These leads were produced by the scoring algorithm and further fine-tuned as most likely to respond. However, reps were suspicious of the potential of prospects with fewer employees. When reps are not completely trusting the sales potential:
Provide them with evidence that shows why these prospects scored high. Share the desirable attributes of these prospects.
As one option, using Industry (SIC) distribution, create “blocks” of leads that contain a mix of company sizes and provide guidance that x% are expected to convert for each of these “blocks.”
A benefit of predictive analytics is the ability to score based on a large number of attributes (500+ variables is not uncommon for our models), reaching beyond the boundaries of even human intuition which no doubt can bring profitable insight. We don’t discount that intuition; rather, predictive analytics should lead you to new sales from new markets. The flip side is that these new markets and new customers may not intuitively look like great sales opportunities at first glance, precisely because they’re new and different.
But acting on model recommendations is needed to break through to faster growth and higher sales. This is where sales managers can mentor reps to see the opportunities ahead. Help sales reps trust the idea of calling upon these new opportunities, and the potential conversion rates and quota attainment, even when reps may have previously rejected some of these opportunities. To get the true gains possible with predictive analytics, you need to build the trust and confidence to push beyond a business-as-usual routine.
On a recent assignment for setting up lead generation, we took on an existing customer database to build a statistical model to score a leads database. The client does business in 7 countries, or so they said and I believed them. But they quickly added a caveat, “no one has looked at our database in awhile.”
First, we looked at their billing country field. This had been an open text field in their Salesforce.com system that could be edited by just about anyone. What we found was amazing. There were so many variations for each country that unique values quickly proliferated to 107. Misspelling, case difference, punctuation and abbreviations all conspired to create many versions of the same country!
Our first order of the day was to identify the obvious countries and group them, followed by corrections to the remaining data. This took 107 countries down to the correct 7. Then the country field was locked and going forward, a drop-down menu of countries is being used to prevent this proliferation from happening again.
For statistical modelers, why is cleaning data important for sales intelligence?
First of all, we want to point out that even a simple field can pose a challenge. In the absence of consistent coding guidance, it becomes hard to create a segmentation or master-filter at the top with which to analyze data. And here, it is not wise to mix multiple countries within the analysis.
Second, there may be other fields with which data can be cross-referenced. If there is a shipping country and a billing country, chances are that both ought to be the same. So correct the data, then carry over the values into the other field so you have fewer blank and invalid rows overall.
Third, the state of data quality can give a clue as to what other fields could be a problem. Even if not used in modeling, always keep in mind how this data was created, who and when it is updated, and what procedures are used to correct bad data.
Finally, proactively start addressing quality and instill an ongoing practice of making data cleanliness everyone’s responsibility. Have a method by which to collect feedback and incorporate it. This way, when you are ready to perform analyses, there are fewer surprises with respect to data quality.
Parting thought: Do not assume anything “should be obvious” or find where to lay blame for bad data. As you can see from the simple example above, all businesses have to approach data hygiene with care, caution and respect. As my friend’s father, an electrician said: “The day I stop fearing electricity is the day I will stop working.”
Your lead generation machine is working tirelessly to bring in new leads and appointments. Marketing keeps hot prospects focused on the message and proposition. Your sales team has done an excellent job of converting them into a first time buyer.
And then something not-so-good happens. The shiny new customer does not come back to make a second purchase. Months go by, and they end up in this no man’s land called “one time buyers.” As a lead, they are coveted. As a buyer they are ironically abandoned. Countless artists have sung the “one and done” blues. Stevie Wonder even has a top hit on this, I think.
True cost of leads
The consequence of underperforming leads is enormous – from metrics such as cost per lead, multi-buyer conversion, average initial sale to cost per buyer. Indications are that a hidden deficiency to turn first-time customers into repeat buyers seriously understates the true cost of customer acquisition. Lead generation should not stop with the initial sale, rather it should cultivate prospects that contribute to maximizing long-term franchise value. This can only happen through continuing repeat purchases and higher level of engagement.
It’s a safe bet that you do not make money on the first sale after subtracting net profit from the cost to acquire the customer. When you amortize the cost of up-front lead generation costs, consider adding another metric so you can measure against cost to make two purchases, three purchases, five purchases and so on. Given the incremental profitability of repeat purchases, it should come as no surprise that the ROI becomes exponential.
What you can do
So how can you reap the benefits of this promise? Identify the symptoms, then interpret the signals and take decisive corrective action so these leads never slip into the no man’s land. Let us discuss each of these stages in detail.
First, identify how many customers have made only one, two or three purchases. Look at this in time intervals of six, twelve and twenty-four months since initial purchase. Also perform a vintage analysis by quarter for the past two years.
Second, look at the purchase details. Were most customers buying just accessories? Were they rush shipments? Were products only in one product category? Only on sale? Online or offline? Are some sales teams more prone to the “one and done” problem than others? By segmenting these buyers, you will begin to get a good idea of what is producing the majority of these customers.
Finally, put in place a campaign or process to touch these customers early. Inquire about their buying experience. Look at an attach that you could drive immediately. Set multiple touches based on the first order amount, profitability of the product purchased, or time since the initial purchase. Offer a value added service that “welcomes to the organization.”
There are many reasons why a customer may not make a second purchase. But a significant reason often is that there is no follow-up. By establishing a proactive process, you can extend the ROI of lead generation significantly through simple steps that resonate with the shiny, new customer.
Lead generation comes in all forms, from the old standby of print and email list rentals, to emerging and evolving online search strategies. For offline lists, the focus of this article, there are two types: active and compiled lists. After explaining the differences between the two, we’ll focus on compiled lists and best practices to improve lead generation results from prospecting efforts.
Active lists reflect recent (i.e., less than 12 months ago) activity such as magazine subscribers, event attendees, responders or buyers. These are often contact-level lists. Buy small samples from each list to test response before investing in a bigger campaign. Check that there is sufficient quantity of names available for a larger rental later, and be sure to take into account any subtractions such as do-not-mails, geographic limitations, etc.
You can select desirable attributes relevant to your targeting needs, such as contact title and decision-making authority. But caution: although selects can produce more targeted prospects, they dramatically lower the available universe of names. Selects may also not be indicative of response until you have tested the list several times.
Compiled lists are a broad compilation of individual or company records. Due to the extensive number of records gathered, you will often have a number of selects to choose from. In this case, you will want to use the appropriate selects to target the best leads for your product or service.
Many selects such geography or industry may be obvious choices. Beyond that, deeper selects — and the various combinations of selects – could have strong correlations to response, but these relationships may be difficult to detect until you test. Or to the contrary, they may lower performance because of unforeseen interactions. The conundrum is, these deeper selects, and combinations of them, have potential for significant increases in response, thus improving your prospecting and lead generation investments.
How can analytics help you overcome these issues? There’s a few steps you can take:
Step 1: “Look Alike” Profile
To improve performance of selects, start with “look alike” models. These models compare your best customers to the general list universe, and they identify attributes that “over-index” – meaning that’s where your best customers are found in higher proportion than the general universe. The results will surface a combination of selects that will yield the maximum response.
These models can typically yield an approximate 20% better lead generation response. While this is a nice upgrade, continue to overlay selects that have traditionally worked for you. Test the model against your existing selection criteria both through back-testing and live in a campaign.
One crucial element of success – as well as risk – is how you define your “best customer.” Is this your most responsive customer to a single campaign? Or the most valuable customer after 12 months? For example, let’s say you choose to profile customers in the top 20% of spend as your best customers. The problem with this definition is it ignores how long it took for those customers to become valuable, and does not distinguish between “one large order and done” customer versus one that took ten years to get there. In short, this has little to do with how you acquired that customer – it brings no understanding to how to improve prospecting performance.
Step 2: Multiple Outcome Comparison
One way to address the risk that you’d be off target with a single “look alike” definition is to create multiple definitions to cover a broad range of outcomes, and build and test several models. Start with outcomes that are immediate, such as leads that downloaded product information or set an appointment. You can also evaluate long-term outcomes such as placement of five orders or spending $5,000 in the first year since acquisition, or customer purchase across three product categories.
Other outcomes could be narrow, such as a prospect downloaded a white paper or requested a quote. You can also try niche outcomes such as highest first order amount, or purchases of specific service, warranty or higher margin product.
Analytics can test all of these long-term and narrower outcomes at the same time, and can tell you the most valuable methods of prospecting.
A successful prospecting campaign should not only generate the best immediate response, but contribute the most to your franchise value over time. While it might seem like an arduous task to evaluate numerous outcomes, this is the very path to high-performance prospecting. While solutions are available including Valgen’s to make this process simple through automatic iterative modeling, with careful planning, you can start by testing a few outcomes manually, and evolve over time to a more intensive analytical process if you want more detailed insights.
Step 3: Response Cohort Analysis
A prospect that responds to your campaign by downloading online product information, requesting a quote, accepting an appointment or even placing an order is only one piece of the puzzle. You must ensure that the value proposition that brought the customer to you in the first place is reinforced, through both the circumstances of the purchase itself and their on-boarding experience with you. That’s where the effects of offer, channel and sales process becomes important.
Going beyond multiple outcome comparison, you also want to capture key information about how prospects are converting to customers. This is what we mean by response cohort – assigning the newly-acquired prospect to a segment containing a unique combination of factors related to the response such as the offer, channel, salesperson’s level of experience and even number of contacts made before first sale. Ensure that you capture this type of data, so you have the ability to refine the quantitative and qualitative aspects of your lead generation process to improve results over time.
After you have settled on the analytic approach to take, it should be paired with the most suitable operational tactics. This will help you validate the analytics, and improve your prospecting processes based on the feedback from your analysis. For example, you may find that with high-scoring prospects, making two additional touches within a month could yield an exponential conversion rate. This finding could only be uncovered by in-depth analytical efforts. We will cover this in another article here soon.
When you change your lead scoring and lead delivery using predictive analytics, don’t forget to train sales reps to think different as well.
Inside sales teams are at it every day, making thousands of calls to prospects, seeking an appointment or a sale. They are given scrubbed lists with contact names, job title, phone numbers and a good luck pat. On the back end, managers track calls, appointments and sales. The cycle continues when reps deplete their assigned leads and it starts over again.
