Not a member? Sample unlocked content here.
Increasing numbers of sales and sales operations teams leverage technology and AI-powered insights to deliver more accurate forecasts, keep deals moving through the pipeline, and focus on the right deals to drive revenue.
In this episode of the social selling and tech show, we will explore how AI and Predictive Forecasting is helping salespeople to get a better handle on their sales leads funnel.
Our guest today is Michael Lock, CEO at Aviso.
Listen to this episode of the CPSA Social Selling and Tech Podcast and discover:
* Why ditch the spreadsheets? How can sales forecasting tech help achieve better results?
* Who needs sales forecasting tech? Is it best suited to specific types of companies?
* How can sales forecasting tools help prevent leads falling through the cracks?
* What can machine learning models tell us about the quality of the pipeline salespeople will need to reach their goals for the next 3, 6 or 12 months?
Want to hear more? Check out these bonus insights:
* How can machine learning allow sales teams to model how deal factors contribute to sales outcomes?
* What are some of the typical user adoption challenges, and can you suggest some ways that these can be overcome?
* Can predictive forecasting tech help salespeople get a clearer picture of the deals that are, perhaps surprisingly, more likely to close?
Read the edited transcription:
Bill Banham: In this episode of the Social Selling and Tech Show we are joined by Michael Lock, CEO at Aviso. Increasing number of sales and sales operations teams leverage tools such as Aviso's technology and AI powered insights to deliver more accurate forecasts, keep deals moving through the pipeline and focus on the right deals to drive revenue. We will explore how AI and predictive forecasting is helping salespeople to get a better handle of their sales leads. Michael Lock welcome to the Social Selling and Tech Show.
Today we are talking about a wonderful and complex topic of AI and predictive forecasting. To start with why ditch the spreadsheets? How can sales forecasting intake help achieve better results?
Michael Lock: Sales forecasting is important to business results in two ways. First, it's very important that everybody knows that you build a sales forecast so that you can communicate both outside the company and inside the company about what the results are likely to be. Most people use the sales forecast to inform their investors, the rest of their community, their board of directors, their executive team about how things are going to look. Inside the company it's used across the whole company then to figure out what you should be doing. Without a proper sales forecast your external stakeholders can get out of alignment with you, and if you're a public company that can bring bad results, and your internal stakeholders can get very much out of alignment.
The more important area on how sales forecasting produces better results is it actually you improve. The problem most often is without proper sales forecasting you don't realize until the end of the month, the end of the quarter, the end of the year that results are going to be different than what you expected, and then you're not taking any actions 'to improve them, so proper sales forecasting, where you identify gaps in what's happening in the business allow you to take action and actually improve the results to do. So it's the second part there that's actually the most important that's there that people forget about and that's why sales forecasting is such a important topic for most business executives.
Bill Banham: Is it best suited to particular types of companies, such as fast growth organizations?
Michael Lock: First of all, being excellent at sales forecasting is important to all companies and that isn't companies in a particular industry or in high growth or in low growth. In the past we used to have businesses that were much more predictable. We have a business and we take GDP and we move it up. If the economy grows by 2%, we're going to grow 2%.
In today's fast changing world there are almost no businesses that are highly predictable in the old way that they were, and so the use of modern technologies, big data, AI, machine learning are going to be necessary in order to help the business predict where it is going. It is widely applicable. We've only started at technology companies because I figured if we can't talk technology companies into using big data and AI in sales then we're not going to be able to talk manufacturing companies into, but I believe that sales forecasting is applicable across a wide set of industries, whether you're a manufacturer, a high tech, whether your sales model is lots of transactions or whether your sales model is very few transactions. All those things are going to need proper sales forecasting and of course powered by technology.
Bill Banham: How can sales forecasting tools help prevent leads falling through the cracks?
Michael Lock: Particular tool starts with sales forecasting and then applies AI across both pipeline and deal review. We can look across the entire number of deal you have within your pipeline, and we can say not only what is the average sales cycle length. We can say, "In stage three it takes this long." Or, "In stage one," since we're talking about sales leads, "It usually takes us eight days to get it out from stage one to stage two." And our technology will then identify and say, "Wait a second this one is taking 15 days to get out of stage one," and therefore something gone wrong. Are we not calling them back? Are we not provided the proper information to the prospect to do that? And that's really across the entire sales cycle.
On the other end of the cycle, if you're at stage six where you're negotiating the contract, if that is taking longer that what is normal then AI and machine learning technologies along with big data can identify that, bring that attention of sales management and sales leadership and action can be taken so that things don't fall through the crack, not only at the lead, but at any stage in the entire sales process.
Bill Banham: Can you paint a bit of a picture of the sales results in companies which combine the use of CRMs with big data, with AI and with machine learning algorithms?
Michael Lock: Yeah. Let me first start with, to paint that picture to talk about how it is generally done today without big data and artificial intelligence. Most companies by now have installed a CRM system, and what that CRM does is it tells, it's much better than when we didn't have our CRM systems, it tells you what is happening with the deals, with the pipeline, with the sales at a single point in time. And so you can look, a sales manager can go into it, or a sales rep, or a sales leader can go in and say, "Hey, this deal is this large, at this stage, at this close date, at this time."
The problem with CRM technology is that it doesn't tell what it looked like last week, or the week before, or last month, and so you're left with human memory going, "Geez I wonder ... I think I remember it being like this," and so what people do is they create a separate database and boy that separate database unbelievably is generally an excel spreadsheet, and they put the history of the deal or the pipeline in there, and people then are stuck looking at these two disconnected and neither of them is providing. The spreadsheet doesn't provide the level of details that's necessary in order to take action and the CRM system doesn't provide the historical view of this, or what we call a time series view of what's happening in order to make that happen.
So what are system is doing is that we're making that one thing. We actually take a snapshot of the CRM every 15 minutes. We put it in a huge cloud-based database. We then run artificial intelligence and machine run algorithms against that to predict the computer helps predict what the aggregate forecast should be, as well as what the forecast is at every business unit level, every business segment level, every product level, every rep level. It's a much more comprehensive way for you to look at your sales forecast. We're not completely relying on the computer, the robotic aspect of this. We take a machine learning forecast and we combine that with a human being forecast, the traditional rep, manager forecast. We compare those two things and we allow then the computer and the humans to kind of shake hands and say, "Here's a set of data points. Let's figured out what's happening with the business. Let's figure out what our sales forecast would be, and also let's figure out what we need to be doing with the pipeline."
Bill Banham: Okay, so let's imagine sales people are using bug data, they're using machine learning, they get it, they're running on all cylinders, they're closing more deals, but wait. What about those gaps where the leads could maybe drip through and be lost? Can predictive forecasting help to identify those gaps in your sales process and the leads funnel?
Michael Lock: Absolutely. What we have done is we've built this very large time series database, so we can see how the deals over time, so we're tracking, "Hey, how long does it take to get from stage one to stage two, from stage three to stage four." We're tracking whether the deal is, you know, what is the size of the deal? What happened with the close date? We're tracking every field and deal attribute that a customer has in their CRM, so we're able to track, hey , what is ... I'll give you two good examples here. Most customers set up a field that says, "Who is the competitor on this deal?"
Another example would be they set up a field to say, "What's the current installed solution that day?" Those are often highly predictive about whether that deal is going to close, how much it's going to close for, how fast it will go, and yet we tend not to have a good read on those on the way we do pipeline analysis, pipeline management, and forecasting today. So by looking at every attribute that exists in the CRM we're able to then tell people in a very factual way what are the things that are most too ... When's a deal mostly likely to close based off historical data, and that is a really big step forward, and customers are often somewhat surprised in what the factors are that contribute to a close sale.
Bill Banham: We're coming towards the end of this particular interview Michael. Before we wrap things up just one last question. To recap, what can machine learning models tell us about the quality of the pipeline to reach their goals for the next three, six or 12 months?
Michael Lock: I think that's a really good question to wrap up on. Traditionally we've had a decent set of tools to tell you about pipeline quantity. You can look almost at any time into your sales force dot com or Microsoft dynamics system and say how good is my pipeline. What traditional systems are not good at is telling what's the quality of that like look, so what machine learning system like Aviso's do, is we say "These are the elements in a deal that make it a good deal, and how many of those deals do you have." If you have a bunch of deals which are against your toughest competitor, those are of lower quality than the ones that are against your weakest competitor. If you have a whole bunch of deals that are in let's just say a certain industry, say hey, you're really good at selling the banking, but most of the deal you have in the pipeline are in healthcare, then you realize the quality based of off the historical pattern of what's closed in the past is potentially not as good.
One of the most important things that big data machine learning can do is it can actually help you figure out what quality is in the pipeline. What is most likely to close, and do I have enough of those deals in my pipeline actually to make my number? Having a bunch of deals in the pipeline that are not likely to close is actually the worst possible thing you can have. It means you're going to spend a lot of time and resource selling to people who are not going to buy, and what we need sometimes is more concise pipelines, more focused pipelines, in industries, against competitors, in installed basis which we're more likely to close, so focusing on quality, machine learning big data is going to be a lot better than human judgment or going by the seat of the pants, which is what we've been doing for the last 25 years.
Bill Banham: And that's a great note to finish on. Ladies and gentlemen, no longer is there any excuse to go by the seat of your pants. Michael Lock thank you very much for being the guest today.
Michael Lock: All right. Thanks.
Already a member? Login to see full the article.
Not a member? Sample unlocked content here.