Outperform Your Competition by Leveraging Artificial Intelligence to Enhance Your Sales and Marketing Funnel

Included Research:

By Shashi Bellamkonda, Principal Research Director

Executive Summary

Marketing teams collect massive amounts of data on prospects and the prospect journey. However, these teams struggle with integrating and analyzing data to prioritize prospects. This leads to missed revenue potential.

This problem is compounded by teams creating analog and transactional lead prioritization, which does not consider any behavioral metrics. Buyer journeys often do not follow a set pattern, and identifying and prioritizing the best prospects is more instinctive and less data driven. Moreover, often there is friction between Marketing and Sales on what is seen as a lead that must be prioritized.

Comparing buyer behavior with profile data and the interactions of converted prospects will give marketing teams an effective way to help Sales prioritize prospects that are most likely to convert. There are many ways artificial intelligence (AI) can be used in marketing to this end. This article will address how your organization can leverage AI as its most critical application to quickly improve revenue at the bottom of the sales funnel.

Adopting this project can serve as a solid foundation for other AI applications in marketing, including:

1. Marketing automation
2. Content generation using AI
3. Chatbots for customer engagement
4. Content search optimization
5. Customer and prospect analysis
6. Personalized marketing campaigns

Challenge the traditional beliefs or norms

Several schools of thought may think that it will be impossible to use AI to increase customer acquisition. Technology constantly changes and using AI or machine learning (ML) can analyze data faster and give a better understanding of prospects with high intent quickly.

If you are on the roadmap toward leveraging AI, you may think that the AI components that exist in your customer relationship management (CRM) system or other software will suffice for your needs. However, to improve bottom-funnel conversions, you should use your data to create a customized AI/ML model that will help identify high-converting prospects quickly and increase revenues. Getting this right over time is especially important to early and mid-stage start-ups whose leaders and investors demand getting to scale. Using this strategy, you will be ahead of the curve, developing faster ways to optimize your ideal customer profile (ICP) and buyer journey in an agile, iterative fashion.

Benefits of creating an AI model to increase down-funnel conversions

A major advantage of this AI/ML model is to speed up sales cycles, reduce cost of sales, and improve marketing affected gains:

“In a survey conducted among global marketers in late 2020, it was found that 41 percent of respondents saw an increase in revenue growth and improved performance owing to the use of AI in their marketing campaigns. Another 38 percent attributed creating personalized consumer experiences to AI use for marketing purposes.”

– Statista, “Outcomes of AI-enabled marketing worldwide 2020”

With proper implementation, we can expect the following benefits

Improved revenue through enhanced conversions: Every increase in conversion rate can result in thousands of dollars in additional revenue. As shown below, a 2% improvement in the MQL-to-SQL conversion for a company generating $17 million in yearly revenue and growing at 35% would result in an increase of $187,200, while a 2% boost in SQL-to-Demo conversion would result in an additional $617,760.

Applying AI/ML to improve bottom-of-funnel conversions gives you measurable benefits as proof points when requesting funding and resources for additional uses of AI/ML in marketing. In a case study by AI-as-a-service provider Abacus.ai, a customer insights company using AI/ML to generate personalized recommendations increased revenue by 20%. Key benefits included:

1. Easy to scale resources: Efficient growth means you can deliver more revenue with existing resources.
2. Efficient usage of the data: The data collected on prospects is put to beneficial use to get more revenue.
3. Single view of the prospect data: If your systems are not currently fully integrated, piping the data to an AI model can give you actionable insights to help power a dashboard.
4. Better align marketing content with the customer acquisition journey.
5. Better enable advanced strategies such as account-based marketing (ABM) and product-led growth.

Why is this important now?

AI uses for marketing is not new. Amazon, Netflix, Google, and Target have been using AI to customize the offering to their customers using machine learning. A survey by McKinsey found that this year 50% of firms across the world had tried to use AI in some way, up from 20% in 2017.

A decade from now, all businesses will be Al businesses – make sure yours makes that transition better than your competitors.”

– Peter Gentsch, AI in Marketing, Sales and Service

“I believe deeply – to my bones – that the most important development in the history of marketing is machine learning…it will fundamentally change our relationship with consumers.”

– Kristin Lemkau, JPMorgan Chase CMO, “AI in Marketing: Benefits, Use Cases, and Examples”

Source: Info Tech Research Group, CIO Priorities 2023

As Info-Tech’s own research shows, 44% of organizations are planning to invest in AI or ML by the end of 2023. Ensure Marketing is at the forefront of leveraging these new investments.

Leaps of technological advancement make it easier for even early and mid-stage companies to use AI/ML

How to get started: Steps to implement an AI/ML model to use in your marketing

“If your brand does not have the resources of a giant network such as Amazon at its disposal or a technology visionary with deep pockets at the helm of your company, do not despair. There are significant gains to be had by implementing AI and machine-learning technology to supercharge your brand’s customer journey by delivering the personalization that consumers now expect, no matter where you are starting.”

– Jim Lecinski and Rajkumar Venkatesan, The AI Marketing Canvas: A Five Stage Roadmap to Implementing Artificial Intelligence in Marketing

The impact of using AI in marketing is growing revenue by finding the leads that closely resemble existing customers speedily and enabling sales to reach out faster. Integrating AI into your marketing and sales funnel is also affordable by using AI as a service (AIaaS) companies that have basic models that can learn from your data and give you actionable steps to increase your sales speed. Here are the steps we recommend:

1. Educate the leadership team: Introducing AI/ML into your marketing is an initiative that will need to be socialized and get buy-in from the leadership. Usually the CEO, COO, CTO, CFO, and CSO engage in these discussions. Use the data in this article or reach out to our team to help with a presentation to present the concept and the benefits.
2. Collaborate and plan with sales and tech: To succeed, this initiative must be co-sponsored by Sales, Marketing, and IT. The project will need support from the tech leaders to assign resources for data collection and transfer and to decide on the build versus buy model. Info-Tech can help your marketing and IT teams flesh out your strategy and get your project focused on accomplishing your business and IT goals.

Source: Info-Tech Research Group, Create an Architecture for AI – Phase 1: Assess Business Use Cases for AI Readiness

3. Decide on buy versus build: Building an AI/ML model in house needs a big budget and skills and knowledge that are usually beyond the reach of small to medium tech companies. Most models once built may not need a full-fledged in-house data scientist just for this project. Also, this technology moves so fast that building in house may make the technology redundant soon.
4. Build requirements – key elements include:
   • Specify the desired output data.
   • Indicate the location for displaying the data.
   • Determine the frequency of AI/ML analysis for leads.
   • Define the retraining intervals for the AI/ML model.
   • Outline the method for transferring data to your CRM or sales enablement tool.
   • Identify the location of the data, whether in a data lake or sourced from Salesforce or another CRM.
   • Define the rating or scoring metrics and the associated scale.
   • Determine the frequency of retraining the AI/ML model
   • Use supervised or unsupervised machine learning. We recommend supervised.
5. Identify potential vendors: SoftwareReviews has a list of AI/ML at Best Machine Learning Tools 2023 | SoftwareReviews and you could talk to an analyst to learn more. Other lists include The AI 50 2022 | Forbes.com and Top AI as a Service Companies | VentureRadar.
6. Complete selection of the solution: You will find this Info-Tech tool, Fast Track Tool: Select and Evaluate Vendors , useful in structuring your AI/ML vendor evaluations.
7. Ask for a proof of concept (PoC): Given that the field of AI is new and companies offering AI as a service are also relatively new, request a PoC from the vendor. This allows you to test the service and confirm that it meets your expectations and produces the desired results. The vendor will provide a trial period during the PoC.
8. Scope pilots properly: Pilot your bottom-funnel use case with the right target ICP and buyer journey so when improvements are measured, results can project upon a larger data set with confidence to assure approvals for rollout.
9. Create training data for the AI model: One of the best surprises is that an AI project is the requirement phase. There will be no need to do detailed research or analysis, as we will use existing customer data for training the AI. A company with “bad” customer data has other problems than just the implementation of AI! The training data needs these steps:
   1. Labeling the data that belongs to customers.
   2. Collecting the data on this data set of customers including all the attributes and interactions.
   3. Deciding what will be the key that is used to identify the individual customer.
      • Training data: 75% of the customer data will be piped into the model to train the model.
      • Testing data: This is the 25% of the data where you know the answer as they are already customers. Once the AI/ML model is trained this training data is given to the model along with other prospect data to test if the machine learning model will correctly identify the 25% data as the prospects most likely to convert.
   
   
10. Evaluate the results: The results of your AI/ML model compared to the reality will always be changing. The advice AI/ML experts give is not to measure the effectiveness of AI because of a test with static data but to measure the outcome of the AI results with a measure of human intervention. Past models have shown 65% to 75% accuracy in test data, which is particularly good.
    According to Obviously AI, “Good accuracy in machine learning is subjective. … but anything greater than 70% is a great model performance. In fact, an accuracy measure of anything between 70%-90% is not only ideal, but also realistic. This is also consistent with industry standards” (“How To Know if Your Machine Learning Model Has Good Performance”).
    Here are example results from an AI/ML model that has trained on a sales and marketing funnel. All the opportunities were analyzed by the model and scored in the column “AI/ML” and the actual results are in the column “AI Success Rate.”

• The way to interpret this data is that the model creates a score of 0 to 1, and for all the deals scored above .5 as a predictor there were 74 actual deals and the deals predicted by the model were around 37; therefore, it had an accuracy of 55%.
• For the top two categories (above .8 score) where the highest number of deals were closed, the model predicted 60% of the deals.

11. Adopt AI: Training in sales and marketing. This is an extremely critical step. Sales teams are focused on sales and must buy into using the AI model results. An approach for convincing them is to train them on the basics of how the AI is built. Make the connection and distinction between the generative AI that they hear about everyday online with your AI/ML model. The sales leader can use a few sales team members to beta test the model against their existing way to prioritize the leads.
12. Create a rough timeline:
    1. Create model and test data: 60 to 90 days (about 3 months).
    2. Observation: 60 to 90 days (about 3 months).
    3. Present results to leadership team.
    4. Kick-off to Sales and Marketing: 7 days.
    5. Implement test plan with a section of the sales team: 30 days (about 4.5 weeks).
    6. Evaluate results and expose the scores to entire sales team: 7 days.
    7. Continue to add more data to retrain the model.
    8. Check results every quarter.
13. Add more data attributes: For marketers who have reached this stage, future steps are to continually improve the model's performance through learning and adding more data attributes, usually enhancing the data through a customer data platform . The model sets up a hierarchy of the attributes' significance. For example:
    1. If a prospect unsubscribes, the model figures out that this prospect has the highest probability of not closing.
    2. The model uses industry data to correlate it with successful customers.
    3. The model can also analyze earlier data to predict that prospects assigned to high-performing team members are more likely to close.
    4. These correlations are based on supervised learning, which requires that the data be tagged and not influenced by notes or other manual entries.

Prepare for Obstacles

Our research shows program leaders will encounter friction and hurdles as you go through your project. Plan to address:

• Data limitations
• Incorrect data
• Insufficient data
• Inaccessible or fragmented data
• Doubts about the accuracy of an AI-generated scoring model
• Resistance to change from the sales team
• Lack of technical support
• Incompatible technology stacks with application programming interfaces (APIs)

Numerous foundational models have already been trained on similar or synthetic data. They may not need vast amounts of data. Your AIaaS vendor will be able to help assess how much data is needed. Our experience has shown that different organizations have different requirements. We have seen companies work with data available with 250 customers, 1,000 customers, or looking in a unique way at about five million records in a data set, which may equate to an early stage to medium software company with a well-organized marketing engine.

Implications of doing nothing

The consequences of postponing or neglecting the importance of making AI/ML a crucial part of your marketing strategy can be severe. Your sales may slide, and your company does not scale to support future growth.

You will lose an opportunity to enable innovative marketing strategies and increase efficiency in processes.

“Do not wait for the marketing world to get smarter around you. “Take the initiative now to understand, pilot, and scale AI.”

– Marketing AI Institute, “Piloting AI in Marketing”

“The opportunities are endless for marketers with the will and vision to transform their careers, and the industry.”

– Paul Roetzer, Founder and CEO (Chief Executive Officer), Marketing AI Institute, “Keynote—Dawn of the Next-Gen B2B Marketer with Paul Roetzer”

Conclusion

The key to success in marketing lies in quickly adopting new tools and automation strategies, including the use of AI. Without using AI to identify the best converting prospects, you will be wasting resources and your growth will be lower than competitors who are using AI in their marketing and sales funnel. Implementing AI in marketing opens the door to incorporating AI into other aspects of marketing and enhancing the customer experience. By using AI and machine learning, organizations can grow and scale faster and be better equipped to weather economic challenges by perfecting their use of time and resources. This benefits the workforce, as they no longer waste time on unproductive prospects, and improves relationships between the growth and demand generation marketing teams. It is a win-win scenario.

Other Related Research From Info-Tech Research Group

Optimize Lead Generation With Lead Scoring

No Data Scientist? – No Problem; You Can Still Get Your Feet Wet With AI!

Drive Business Value With Off-the-Shelf AI

Get Your Feet Wet in Cognitive Analytics With AI-as-a-Service