Marketing: Why won’t the salespeople work our leads?
Sales: Why won’t the marketing people give us good leads?
Sound familiar? Part of the answer to this age-old dilemma is lead scoring. In Part 1 of this series, we looked at the benefits of lead scoring and discussed the data that is needed in order to score leads. In this post, let’s dive into the different options for lead scoring.
Before You Start
Yep, there’s always something else you have to do before you start. In this case, you should do some analysis and collaboration before you try to setup a scoring system. Analyze your current list of customers to see if you can identify what a typical customer looks like. Better yet, try to differentiate them from the standard marketplace. For example, you may find that 15% of your customers are in healthcare and that is your largest single industry; and you may find that professional services is only 3% of your customer base. So you may be inclined to score healthcare prospects higher than professional services. But if you analyze your lead conversion history, you may find that 70% of your leads were healthcare and only 1% were professional services. In other words, you have a much higher chance of converting a professional services lead into a customer than you do a healthcare lead.
Once you think you have a good target profile for a lead, collaborate with your sales and marketing teams to see what they think. Sometimes they will be able to tell you something that the data doesn’t, such as “we hired a guy 6 months ago who only works with professional services firms and his old company went out of business so he brought all of his clients with him, so it makes sense that we have a lot of them, but that doesn’t make them a good prospect for us.” The objective should be to either have the analyst explain the data to them so that they understand and accept it; or that they explain what is wrong with the data to the analyst such that they refine the analysis until it aligns with what they say. Note that I did not say, “throw out the analysis and just do what sales and marketing says,” nor did I say, “ignore sales and marketing and just trust the numbers.”
Okay, this isn’t really scoring at all. But it is a decent poor man’s substitute for scoring. Most customer relationship management (CRM) tools these days have powerful filtering features built right in. If you’re a small shop with a tight budget, you can start by filtering your leads down using these features. If your organization has a lot of clients in the bead making industry with under $1 million in revenues, for example, start by filtering on that.
The next option is to use rules to score your leads. Rules include manually developed formulas to score your leads. An example formula may look like this:
- If the lead came from the website, give it a score of 10
- If they said that they are making a buying decision within 3 months, add another 10 points
- If they said the buying decision was in 3-12 months, add only 5 points
- If their last contact with us was less than 30 days ago, add 10
- If it was 30-60 days ago, add 5
- If it was over 365 days ago, subtract 10
The list of rules can go on for as long as you like, but you get the point. The upside of rules-based scoring is that it is generally not too difficult to do as long as you keep it pretty simple. The downside is that it is also not usually very scientific, although it should certainly be better than filtering.
Scoring Web Traffic
This is really the same as rules-based scores. The difference is that many email/web marketing tools that integrate with CRM platforms have built in scoring algorithms that you can tune yourself. Web visitation patterns can be a very powerful purchasing predictor. For example, someone who came to your webinar may look like a great lead. But someone who clicked on the invitation, filled out the form, spent an hour on your blog site, downloaded a white paper, but didn’t actually attend the webinar may be a far better lead. If you are using Dynamics CRM or Salesforce.com, there are many outstanding marketing tools available as add-ons. If you’re serious about your marketing, you really owe it to yourself to integrate one of these with your CRM system.
Break out the propeller caps and the pocket protectors, we’re diving into the deep end!
Predictive models are statistical tools used to predict a specific outcome. In our case, they would be used to predict the likelihood of a lead to be qualified and to convert into business for your organization.
There are some significant upsides. They are very scientific and much more accurate than any of the other methods mentioned above. You also don’t have to make up the “rules” on your own. Rather, you run two sets of data through the predictive model – one set is a list of prospects who became customers, the other set would be a list of prospects who did not. The predictive model automatically determines what differentiates these two groups and produces a formula that you can plug in to calculate a score.
The downsides can also be daunting. You need quite a bit of data to create enough statistical significance for most predictive models to work. It can also take quite a bit of time to prepare all of the data for a predictive model. You’ll ideally want some of that customer buying behavior and RFM data that I mentioned in part 1 of this series. You will need statistical expertise to aid you with developing your models too, and the right statistical software. Lastly, to integrate the output of these models into your CRM system, you may need some fairly sophisticated development work.
To summarize the pros and cons: predictive models are powerful, but they are also fairly costly. If you have a large database and an improvement in your lead conversion rate can spell a significant bottom-line impact, then you should consider this approach.
As an aside, predictive models come in a lot of different varieties. Univariate, regression, tree or CHAID-CART, and neural networks are all different varieties of predictive modeling. There are different situations under which you would use each and different costs.
Not quite in the same category as the other scoring tools above, but worth a brief note. Collaborative filtering is a scoring mechanism used to recommend a item to add to your purchase. Think of going through a drive-through and being asked, “since you’re already ordering a burger and a soda to shorten your life by 6 months, would you like to shave a few more days off by getting some fries with that?” Only without as much wit and on a website. E-commerce sites regularly use this type of scoring so when you’re buying that juicer, they also recommend a recipe book and a box of Pepto-Bismol.
This approach can be relevant in a few situations. If you have an e-commerce element to your business, the integrating collaborative filtering into your e-commerce site can significantly boost per-purchase revenues. Some businesses have also experienced benefit by integrating this into their in-person sales process as well. So, for example, if you are discussing a specific product with a prospect, a list of related products can be presented for further discussion. This works well in transactional sales processes (i.e. "would you like to buy some beer since you're already buying diapers?") but not so much with large transactions or strategic sales processes (i.e. "would you like to purchase a bulldozer with that curbing equipment?").
Now that you’ve scored up your list of leads, your next step is to sort them. Create the right view of the data so that when the sales team starts to tackle them, the hottest leads will be at the top of the list. And continue to collaborate with them knowing that your scoring system isn’t perfect, so you’ll want to fine-tune it.
Coming up in Part 3 of this series, we’ll discuss some ways you can collaborate with your sales team to further improve their effectiveness in working the scored lead list.
For more information about this blog entry or C5 Insight, please Contact Us.