Sales forecasting is one of those business activities that is more art than science. Predicting future sales is tough as future results will be affected by:
- Macro-economic trends (e.g., GDP changes, currency fluctuations, etc.)
- Competitor pricing changes
- Changes in one’s own sales team, territories, etc.
- Changes in one’s product mix and pricing
- Fluid customer tastes
- Changes in fashion trends
- New competitor entrants
- New or substitute solutions being introduced
- Changing alliances with key channel partners
Sales professionals and chief revenue officers who are tasked with long-range forecasts fond those same forecasts come with increasing difficulty and imprecision. How can someone predict sales 18 months from now when the average deal originates and closes in mere days or weeks?
Despite their ubiquity, spreadsheets won’t cut it in most medium to large firms. And, a hodgepodge of CRM, financials, planning/forecasting and other tools aren’t exactly heaven-sent either. Businesses need a better way to forecast sales/revenues. Can any of the newer machine learning/big data/etc. tools make this process any better or accurate?
Dave Kellogg, CEO of CPM vendor Host Analytics, has described how planning has improved over the years. He shows an interesting timeline where business systems used to only report what could be measured. Now, with new kinds of technologies and data sources, companies have a real chance of measuring what needs to be measured. The difference is important and lies at the heart of what must be attended to in sales forecasting.
A new generation of sales management/forecasting
Developing a great, accurate sales forecast may involve the use of information from internal systems like:
- CRM (customer relationship management)
- SFA (sales force automation)
- Sales Compensation/Commission Tracking
- Quota Management
- Territory Management
- Order Management
- Financial Applications
Additionally, the company may benefit from the use of external data sources like:
- Econometric data (e.g., GDP growth, currency exchange rates, employment statistics, etc.)
- Competitor sales & pricing data
- Social sentiment data
- Weather data
New vendors are emerging that utilize new data sources, algorithms, in-memory processing and more to get a keener insight into future sales.
One product I recently reviewed showed how using sophisticated time-series analysis of sales professionals’ past deals and past estimates of future deal closures could be used to gain a more accurate estimation of future deal activity. I liked that this solution could determine which sales professionals are unintentionally under/over optimistic about closing deals or when those deals might actually close.
It also looked at other factors to determine what kinds of deals a person has little trouble closing (e.g., smaller deals from long-standing customers) versus those that have lower closing probability (e.g., a net-new prospect with a whale-sale opportunity). This technology has one drawback though: it works best with stable sales teams and territories. High turnover, volatile environments may not get much benefit from this technology.
Optymyze is another player in this space but with a different set of genes. According to Optymyze, they:
(revolutionize) the performance of salespeople with data science, process automation, enterprise planning, and advanced analytics. We do this by motivating people to achieve sales goals; efficiently managing sales operations; predicting the performance of each sales person, and forecasting sales results and the cost of sales.
Optymyze offers sales operations as a service. Think of it as a collection of sales automation tools coupled with powerful analytics. Their solution is designed to ensure that ALL aspects of the selling process are ALL working well together and aligned to deliver maximum (and transparent) results.
Like some of the IIoT (industrial internet of things) analytic solution providers, each customer’s solution environment is different and the best solutions are often optimized for each individual customer. In practical terms, this means these solutions are often a combination of unique technologies, some reusable components, and, a mix of standard and one-off processes and KPIs.
These solutions are not a generic, one-size-fits-all kind of toolset like an ERP product. As a result, customers should expect their solution and data may reside in a semi-unique cloud environment but not in a multi-tenant fashion. It will likely resemble a managed services solution instead.
Analytics are often very ‘personal’. What works for one customer may not for another. The factors that determine the likelihood of one firm’s successful sales in one accounting period may be quite different to those of a competitor. Different firms have different customer demographics, different commission programs and incentives, different histories with their buyers, etc. To understand future sales, one must look at these differences – a generic tool won’t work.
The best analytic solution providers possess libraries or app stores where a number of integrations, KPIs, dashboards and algorithms can be re-used. Yes, some of these components are not relevant for some companies/customers and many may require tailoring based on the uniqueness of the customer.
In Sales, one should expect to find wide variations in the underlying commission accounting, CRM, etc. systems. Some of these systems may be sophisticated package solutions and others no more than a spreadsheet. Yet an analytic solution provider must find a way to work with all of this variety. Optymyze has already developed a number of integrations with common ERP, CRM and other applications. Some of these integrations occur in real-time.
Optymyze also has approximately 200 industry-specific solutions within its App Gallery. Customers and partners can develop these solutions and make them available to others via this app store. Customers can also make functionality and UI adjustments via Optymyze’s app studio.
I’m particularly impressed with their analytics. Optymyze can use dozens or hundreds of data sources to apply its algorithms. These algorithms are used to better understand what drives great sales performance and future results. In some scenarios, the company can help customers reduce overall sales costs.
Optymyze uses Amazon AWS for its cloud hosting/utility computing environment. Several pieces of Oracle technology power the solution. Other Optymyze partners include MapLarge (to provide map overlays to understand sales territories), HighCharts, Logi and Docker.
Optymyze is not alone in its goal of helping sales organizations succeed. There are vendors that offer basic sales automation, sales compensation and other solutions. Moving more up-market, buyers can find sophisticated solutions that aid in the management of sales people, finances, sales operations and more. And, at the top of the pile, one can now find firms that offer intense and insightful analytics. The latter category of solutions can have the effect of eliminating some sales manager functions (e.g., sales forecasting) and replacing inaccurate guesstimates with more real-time planning and modeling capabilities.
Anaplan is clearly gunning for this space as are traditional competitors like Xactly, CallidusCloud, Nice and IBM. Anaplan is rolling out a platform, expanding channel partner relationships and encouraging development of companion applications to work with its products.
Optymyze customers include well-known firms like Office Depot and McKesson. Their ideal customer profile would be a firm with 200 or more sales professionals. These firms would likely have large, complex sales and reside in industries like insurance, telecommunications, high tech, pharmaceuticals and wholesale distribution.
Estimating/Forecasting top-line revenue is one of the most critical budgeting numbers a company must have. Why? Many costs (e.g., cost of goods sold, raw materials purchases, labor costs, etc.) are all dependent on the dollars and mix of goods/solutions being sold. If you get the top-line number wrong, the rest of the budget is worthless.
Worse, an overly optimistic top-line number could be very expensive as sales won’t meet expectations and costs will be too high. Alternatively, if the top-line estimate is too low, then the company may miss out in capturing new market share. More accurate top-line forecasts are key to delighting (not disappointing) shareholders.
I like to see the use of AI/ML in developing ever more accurate sales forecasts. But, I really like the solutions that also incorporate external data points as rehashing internally generated data alone can only drive marginally better results (and only to a certain extent).
Perfection in sales forecasts will always be a dream but greater accuracy is a possibility. If a better forecast could help your firm (and many firms would agree with this), then a better set of forecasting tools (beyond spreadsheets) is definitely needed.
Image credit - via vendor