Is this the future of sales analytics?

MyCustomer.com
Share this content

Even though sales analytics tools have been around for years in one shape or form, most organisations still tend to use Excel spreadsheets to manage their pipelines and help them understand what has happened in the past.

This is despite the fact that the majority of salesforce automation (SFA) or enterprise resource planning (ERP) systems come with their own perfectly respectable so-called descriptive and diagnostic analytics tools.

Just to clarify what these terms mean, according to Gartner, there are four key types of tools on the market, each of which has a slightly different role and purpose. These are:

  • Descriptive – what happened? For example, which key opportunities did we win or lose?
  • Diagnostic – why did it happen? For instance, we won or lost due to competitive pricing.
  • Predictive – what will happen? For example, win rates are likely to improve next quarter.
  • Prescriptive – how can we make it happen? For instance, we will improve win rates due to more effective value selling.

Spreadsheets fall into the first category and so are fine for looking at historical trends and exploring internal sales workflow and pipeline management issues such as conversion or close rates.

But if you want to look for patterns in more external, client-centred data in order to undertake forecasting and explore issues such as which customer segments are particularly price-sensitive so that action can be taken to encourage a future sale, then predictive analytics is more likely to be your thing.

While each approach has its place, the problem with the descriptive and diagnostic approach, warns Mike Grigsby, vice president of analytics at Omnicom’s strategic communications agency, Target Base, is that “a plateau is reached very quickly as you don’t go beyond the past - and there’s no insight into how to make things better in the future”.

Therefore, he believes that predictive analytics tools are the way forward, even though the market for such products is currently immature, fragmented and in flux, with lots of small players around that are likely to be snapped up by bigger ones.

As for uptake to date, this is highest in sectors such as high tech and consumer packaged goods. But adoption is also picking up in financial and business services as well as healthcare – although Gartner expects it to be another three to four years before the technology really moves into the mainstream.

Clean, accurate and consistent data

Current product options, meanwhile, include standard tools from generalist business intelligence vendors such as Cognos and software that enables organisations to build their own predictive models from scratch by suppliers such as Microsoft.

Other possibilities comprise software-as-a-service-based specialists such as Model N, which sell pre-built models and industry-specific applications that cover either one element of the sales process or come as suites handling three or four.

But a key consideration if opting to go down the predictive analytics route is that, in order to operate effectively, such tools need to work against clean, accurate and consistent data provided by in-house SFA and ERP systems.

The issue today though is that many companies simply do not have faith in the quality of their data – which is why they often bypass such systems entirely and use spreadsheets instead. Another downside is that it can take time to de-duplicate and cleanse this information to ensure that it is in workable order.

As Bhavish Sood, a research director at Gartner, points out: “It’s about maturity. So if you’ve not been doing sales analytics for long, the historical data on which to start won’t be there. As a result, before you start using predictive capabilities, you need to have a culture of using standard business intelligence tools already. If you don’t, you’ll struggle.”

Cesar Brea, chief executive of multichannel marketing consultancy, Force Five Partners, agrees. “You really need enough data to reach statistically meaningful conclusions. So if your system is handling hundreds of leads, it should be fine, but if there’s sub-100, it starts to get hard to make a reasonable judgement.”

But Surya Mukherjee, market researcher Ovum’s senior analyst for information management, believes that it is not just a matter of clean data. In many instances, a cultural shift is also required.

“You might be able to do some basic predictive stuff with the help of new technology, but you’ll really need to move up the maturity curve in terms of your IT and business processes to take real advantage of it,” he says.

For example, many companies today operate a silo mentality, which results in functional heads holding their cards close to their chest and refusing to share their true budgets and plans with colleagues.

Cultural shift

But to truly get the most out of predictive analytics implementations, the focus has to be much more on collaboration, with people sharing data across the organisation in order to work towards the common goal of higher revenues and profitability.

As a for-instance, Mukherjee points out that finance has never really understood how sales people operate because, while they like to deal in hard numbers, sales professionals tend to hedge their bets and say ‘there’s a 50% chance of this or that deal closing’.

“But what technology helps you do is to show finance what the pipeline looks like so they can see that the company will make these sales and profits,” Mukherjee says. “It’s also useful for operations to ensure enough raw materials are available to make the products required and for logistics to ensure there’s enough warehouse space. It’s basically doing what analytics has always promised, but never delivered.”

To make it happen, however, requires a huge change management project supported by the highest levels of management. As a result, says Mukherjee, it is important to start small. This means you should:

  • Identify what kind of use case will have the biggest impact.
  • Build a business case around it.
  • Test it out in workgroups that are receptive to change.
  • Broaden out the scope of the initiative over time.

Force Five Partner’s Brea also believes that putting no more than a year’s worth of relevant data into the analytics system is enough initially as trying to do too much in a bid to get more precise predictions invariably leads to trouble.

“The objective is to come up with a better result than a guess from a sales executive, but don’t ask too much at first. Think of it as an evolving standard,” he says. “So maybe get it 80% as good in the first three months, matching in the first six months and exceeding them in 12 months. It’s about delivering early and often and focusing on improving accuracy rather than on one-time standards.”

As for the last category in Gartner’s model, that of prescriptive analytics, widespread adoption, it seems, is even further out. The tools themselves are only just starting to emerge and are today better suited to business intelligence experts and data scientists than the average sales professional.

In fact, according to Gartner’s Hype Cycle of Emerging Technology, the technology is unlikely to hit the mainstream for at least another five to 10 years – although judging by the work required to make predictive analytics happen, it appears that sales organisations will have quite enough to keep them occupied in the meantime.

About Cath Everett

Replies

Please login or register to join the discussion.

13th Sep 2015 09:38

Great article... Only companies that have a culture of data driven decision making will be successful in implementing predictive analytics.

Thanks (0)
14th Sep 2015 11:27

Thanks Brijesh. Very true, but a difficult thing to create. Especially when your business is large and has legacy. Does Starwood have this type of culture, would you say? If so, what did they do to make that happen?

Thanks (0)
14th Sep 2015 11:28

Thanks Chris... From what I have seen; legacy, sheer scale of the companies both in size and geography, cultural nuances, generational gap in info generation & consumption and so forth makes data driven decisions real hard. Also, hospitality being a peoples business and makes it even harder. I would not say that we are perfect but we have made some great strides into this by incorporating baby steps over the years and then getting more and more data driven... This approach has worked well and I guess it's all about creating a win-win situation that moves the organization and the associates forward. just making sure that we not only create tools for predicting the company performance but also the employees so that both can make adjustments when needed...

Thanks (0)