What are the common forecasting model pitfalls in Sales Operations?

Poor CRM data quality has huge ripple effect w/r to your forecast. Below a few quick wins for Salesforce customers to improve your data quality: 1. Define & implement stage exit criteria - document, use field dependencies, and train your team on it! 2. Track field history of the key fields that go into your forecast, deal signals, and pipeline inspection. 3. Weather it's list views, reports, dashboards - create CRM experiences by personas you're enabling (SDRs, AE, CSMs, Managers, Leaders) and get them what they need. 4. Automate activity tracking Use a solution that gives you reporting access for email, meetings, calls, contacts, and rightly attributes them to opportunities so you actually can use it a signals for automations.

1. Overreliance on historical data 📉: Using only past sales data can lead to inaccuracies. 2. Ignoring external factors 🌍: Consider market conditions, competition, and seasonality. 3. Inadequate data cleaning 🧹: Properly clean and preprocess data for accurate forecasts. 4. Lack of sales team input 👥: Collaborate with sales team for frontline knowledge. 5. Not adjusting for anomalies 🚫: Account for anomalies and outliers in data. 6. Failure to update model regularly ⏰: Continuously update and refine models. 7. Lack of validation and testing 🧪: Validate models against actual sales data. 8. Unrealistic expectations 🌟: Set achievable targets based on accurate forecasts.

Completely agree with the focus on data quality. Ensuring tidy, accurate inputs from the sales team is foundational. This bottom-up approach is vital to capture the granular, on-the-ground data. Equally critical is a robust top-down perspective that encapsulates historical data, macro influences, and seasonality. It’s about creating a holistic picture that’s both detailed and broad, blending the immediate, tactical sales data with the strategic, long-term trends and influences. This synergy of detailed internal data and broad external factors provides a balanced, comprehensive view for more reliable, actionable forecasting.

Embrace a mindset of perpetual enhancement to ensure that your data remains accurate and reliable, reducing the risk of forecasting errors and the need for extensive data cleanup efforts.

In addition to data quality and model improvement/alignment/selection... 1. Inflexibility: Rigid models are problematic. Sales operations often face dynamic environments, and the forecasting model should be flexible enough to accommodate these changes. 2. Over-reliance on Automation: While automation can streamline the forecasting process, over-reliance on automated systems without human oversight can be risk. 3. Lack of Cross-Functional Collaboration: it is critically important to involve other departments such as marketing, finance, and operations in the forecasting process. Failing to collaborate can lead to inaccurate forecasts and missed opportunities.

The model selection depends highly on your GTM motion. Below a few dimensions that influence the selection of the right model: - Sales cycle length - Pricing - Products - Territories Once you know what game you're playing, the selection of the right approach to forecasting becomes a lot easier. -

Ensemble forecasting techniques that combine multiple models to yield more accurate and stable predictions reduces the impact of choosing a single, potentially suboptimal model.

One size does not fit all when it comes to forecasting. As is the case with any initiative, you need to consider the desired outcomes and objectives. A long term forecast for planning is going to require a different model from an in quarter forecast to land on a specific revenue number, for example. Consider what you are trying to achieve, then create a model that works to provide the desired information.

Forecasting models are a tool used by business leaders to predict sales, revenue, leads, new customers, supply and demand, and other key activities. The data generated by the forecasting model is used to define annual targets. These projections would be arbitrary in the absence of a forecasting model, which might cause the team to fall short of its objectives. There are some of these popular forecasting models below. - Length of Sales Cycle Forecasting Model - Time Series Forecasting Model - Demand Forecasting Model - Regression Forecasting Model - Seasonal Forecasting Model While all of them can't be used for sales on their own, you use the information you gather from them to help inform your sales forecast.

Overseeing a talented VP of Sales, we've tackled the issue of model selection head-on to optimize our forecasting. While it's tempting to go for complex models that promise high levels of accuracy, we've found that sometimes simpler models are more effective, especially when they allow for easier interpretation and quicker decision-making. The VP was instrumental in rolling this out, consider the trade-offs between model complexity and interpretability, as well as the model's adaptability to our rapidly changing market conditions It's a collaborative process that goes beyond just picking a model off the shelf; it's about aligning that model closely with our sales strategy, market dynamics, and most importantly, the quality of data we have

The biggest challenge with a sales forecasting model is getting the individual sellers bought in on the math - whether perceived too optimistic or pessimistic. The best way to get a seller bought in is to show them how their historical pipeline converted into closed won revenue based on the drivers of the forecasting model. Tying actuals to projections is key to getting an individual seller bought in.

I'm a big fan of combining different forecast models to validate the accuracy of the model. Why limit yourself to one model when you can combine various bottom-up techniques without creating much overhead. Specifically, compare how you're pacing with forecast categories, forecast calls, forecast prediction and forecast ranges. Ex-post, compare forecast submission to understand if there are patterns by person submitting the forecast.

Encourage team members to critically question model assumptions and outcomes. Promote open discussions on potential sources of bias or error.

Always be iterating. This is a common mantra for Sales Operations where we should always be inspecting, improving, and iterating. Forecast models are no different. Because it's a forecast, there are going to be assumptions made. Test those assumptions, conduct "what if" scenario analysis, and look for ways to continue to improve the accuracy of your forecast. The other recommendation I would make is that as you develop improvements in your model, continue to identify areas where data, process, or other changes can be made or captured to improve the accuracy of your forecast.

Talking about the physical products that you've handling. Forecasters tend to disregard not only market trends, but also product life cycles and competitors product upgrades.

Communicating how the actual forecast model works, and what its shortcoming are is paramount to ensure stakeholders trust the forecast. A few things that typically help: 1. Clear documentation of definitions and processes 2. Explicit agreement on the definitions 3. Context in dashboards/ tools themself (tooltips, explanation, links to documentation) 4. Training, training, training - yes, also for executives and the board!

My Vp of sales, learned clear communication with my team is vital for forecast adoption. In the past, I just sent spreadsheets, assuming they'd understand the model. But gaps in their knowledge led to mistrust in the numbers. Now I take time to explain the methodology simply using visuals and examples. I highlight assumptions and data sources so they know the projections' limitations. Interactivity is key—I encourage questions and input to improve buy-in. Taking a consultative approach, instead of just presenting forecasts, boosts understanding. My team is more confident in the targets when they grasp the logic behind them. Investing in transparent, two-way communication with stakeholders has driven greater forecast accuracy and execution.

In my role, I've seen how poor model communication can jeopardize a portfolio company's success. Misunderstandings stemming from unclear data presentation led to poor strategic decisions. To mitigate this, we revamped our communication strategy. We created user-friendly dashboards and used clear language to convey key insights and limitations from our forecasts. The result? Improved strategic discussions and better resource allocation. In hindsight, effective model communication could have averted costly missteps for a portfolio company and drastically improved its trajectory.

How many times have you had to explain the results from an analysis or model because the audience did not understand the data or have sufficient context? Or others questioned the accuracy of your information? Of your data? Of your results? Forecasting is no different. When presenting/ sharing the information, be sure to communicate the what and the what not (!) of the information and include assumptions and the caveats of the model.

If the forecasting is top down the problem is accountability. There needs to be ownership and accountability. This can only be effective if the forecasting is worked bottom up. Not just on units sales prices. But also on Normalised profit in $$ and ٪٪. Which can be allocated to group and or market segments.

Alignment between the model and the goals is a must have. Don't create models just to have models! :-) You want to forecast the business to show how the business is tracking and performing against its targets. And what action(s) needs to take place to help improve the likelihood of achievement of those targets. Or if you are forecasting for planning purposes, for example for components for a hardware build, you want to align the model to provide the business with the right level of detail to ensure you have enough supply. Align the model with what you are trying to achieve (i.e. business outcomes!).

Integrate Forecasting into Strategic Planning: Make the forecasting process an integral part of strategic planning and decision-making. Ensure that the forecasts are not just a number-crunching exercise but are closely tied to the business goals, strategies, and action plans. This ensures that the forecasts have real-world relevance and are actively used to guide business decisions.

Always be testing, iterating, improving. As ops professionals, we know that the business is constantly changing and evolving; therefore, not unlike other areas of our responsibility, forecasting models need to be tested, updated, and improved for accuracy and efficiency, but also to align with the business strategy and goals.

Like every recurring process, forecasting needs constant iteration and improvement. Most companies operate in fast-changing environments where past assumptions just don't work anymore.

One thing that always needs continuous monitoring, especially at a start up is: how are you selling your service? What is the model that you are proposing to your clients? Is the model working? What is the feedback from your leads? What can be changed? What is something that we thought wouldn't be possible but if we think hard enough we can create something along those lines? Sometimes, something that makes sense for the scientists and experts inside a company, might not make sense for our clients. We need to always be on the lookout to align our company expertise to client needs.

Here is how you can improve the sales forcasting model: 1. Clean the historical data. 2. Start with a simple model. 3. Implement a sales pipeline action plan. 4. Use forecasting tools. 5. Incorporate 'what ifs' and qualitative data. 6. Consider seasonality as a factor in sales forecasting. 7. Encourage collaboration between all departments. 8. Incorporate external data where appropriate. 9. Consider market trends and competition. 10. Hope for the best, and prepare for the worst.

Saw this first hand at Hitachi. Data forecasting models became out dated as the IT Industry changed. Monitoring out dated solutions is no longer relevant. The model should be evergreen, ever changing with the needs of the business as it grows.

You need to have a top-down and bottom-up POV. Managers should be getting forecasts from their direct reports. Manager works with direct reports to refine. Manager makes "team forecast call". Manager shares forecast with reporting manager / sales leader, etc. Every step should get input, confirmation, and agreement both up and down the chain of command. Forecast accuracy should also be tracked for each individual throughout the chain of command.

Demand Estimation is complex exercise and needs a detailed analysis of multiple factors. A bottoms up approach is the most effective. Market Estimates, Sales Trends, Competitive Activity, Seasonality Index are the key factors that need to be considered. Market Visits are important to gather these inputs. The Demand from the market level needs to be aggregated at a Sales Unit level (Branch, Region). The Senior Management then needs to factor in Reserve Stock Requirements (Pipeline Stock) based on lead time for production, Time in Transit from manufacture to distribution warehouse and then to the point of consumption. There is no substitute for these basics. If this foundation is strong the rest can fall in place easily.

What-if scenarios are the superpower. It's not enough to get accurate forecasts; you need to make assumptions about what to do to improve that forecast and then test these assumptions. For example, if you see that in a few months, a revenue gap will appear or you will not hit your revenue goals, you must understand which steps you can take to prevent that and improve the forecast. You can do all this using what-if scenarios and testing several hypothesis

A crucial—and often excruciating—aspect of getting ready for the next fiscal year and tracking sales targets is sales forecasting. Leaders who are unable to see into the future with a crystal ball must instead forecast future sales by evaluating quantitative and occasionally qualitative data.

Once you have finalised your preparation and you can but the budget details to bed. I have always found it valuable to consider the total working expenditure at 80% of the budget profit. This allows for a working buffer. The buffer can be utilised towards development.