In my previous work experience I was in charge of generating more leads for a sales team. We had a (long) list of companies to target. We would research contacts for these accounts, and send emails praying someone would answer positively. If we didn’t hit our targets, we would simply contact more companies. We know most demand gen teams can relate to this.
The approach is hardly ever relevant, yields a high bounce rate, the sales team was not convinced, and more importantly, the targeted companies were going through terrible customer experience and yes, some occasionally expressed their lack of delight in the process.
At MadKudu, we’re committed to making customer journey relevant at scale and that includes changing the way we do outbound. In this article, we’ll go over a play we use to ensure our outbound stands out and thus is high performing. More precisely we’ll cover how we detect account-level intent, and leverage it to create relevance at scale.
G2 Crowd is a software to compare company business software and services based on user ratings and social data. It helps you assess what is best for your business. One of their options allows you to know which companies visit your page or your competitor page on their review site. They give you a lot of information about the visitors, like the domain, the page visited, etc.
You can leverage this information to crush your outbound goals. We are going to explain to you how to use them by explaining the process step by step.
Make an API call every 10 minutes with G2crowd, you can use google scheduler. This code will allow you to search for visitors for the last 10 minutes. For the next steps, we need to get the domain and the page visited from the G2 Crowd API
Every company is not equal. We have to determine quickly which one is relevant or not. In our case, we use the MadKudu API (disclaimer: it’s our API).
Here a sample of what you get. Targeting the most relevant company is important, implement a filter to only continue the process for companies considered as at least good.
Check if this company is already in your CRM. The domain will be used. If the company already exists, then check if it’s a customer or there is already an opportunity opened. We don’t want to contact a current customer or a prospect we are talking to.
Use your tool to find the right prospect. According to your ICP, you will target a different job title. Implement a waterfall process. First, try to find the most relevant prospect with a specific job title. If no one is found, then find the second most relevant prospect with a specific job, and etc
Here an example.
4th step bis
Once you find a contact in the waterfall, make sure you didn’t contact him for the last few weeks.
When you have a new contact, add it into your CRM through the API.
Ping your sales through Zapier
Synchronize your outbound sales tool automatically with your CRM. For example, New leads created into Salesforce will flow automatically into it, and it will trigger an email/sequence.
On this chart, you can see the different steps to make your outbound process stands out and high performing
Of course, If you’re running an Outbound campaign that is successful on your side, let us know what’s working! If you want to learn more, hit us up here
For as long as they’ve existed, SaaS companies all around the world have struggled to optimize their Facebook ad campaigns. Due to a few limitations, they’ve had a hard time fully leveraging Facebook ads in their business. Part of this is because Facebook only keeps data from the past 28 days, which means purchase data from sales cycles that take longer than 28 days can’t be fed back into Facebook’s AI. This has proven to be a big roadblock for many SaaS and B2B companies.
Typically, most SaaS companies bid low on a large audience because they’re unable to properly identify and optimize for high-quality leads. This is a key problem. However, companies that are taking a different approach with their ad strategy are seeing a much higher return on investment. In order to effectively retarget top quality leads on Facebook, you first need to optimize your bidding against leads that you actually want. To accomplish this, we use our partner company Madkudu’s AI. Madkudu’s predictive score identifies a lead’s value at the top of the funnel and is excellent at predicting the amount that a lead will spend based on historical data.
Lead scoring is the present and the future.
Focus on your highest quality leads
Once you’ve started using lead scoring, you’ll want to turn all of your marketing efforts toward your high quality leads. Instead of wasting your time on all of the leads you generate, it’s much more beneficial to just focus on retargeting your top quality leads. To retarget top quality leads, you’ll first want to build a custom audience on Facebook.
If you already have audiences, click the Create Audience dropdown and select Custom Audience.
If you don’t have any audiences, you’ll see audience creation buttons, rather than dropdowns.
Choose Website traffic
Set a rule section. (You have to start with an inclusive one.) There are standard rules and pixel event rules. You’ll need to choose “pixel event rules” and then “purchase” (sometimes you’ll need to type this in if it doesn’t auto-populate).
Select “refine by” and then “aggregated value”. Choose “average of”, “value” greater than (or equal to), and then type in the expected value assigned to good and very good leads (MadKudu can help you get this number if you don’t know it already!).
Give your audience a name (and description, if desired).
Click Create Audience.
When we finish creating your audience, select it during ad set creation to reach the people in it with ads.
However, when you’re creating your custom audience you’ll need to choose to build an audience from a Facebook event. This event is called a “purchase” on Facebook, even though it just represents “lead score.” To take it a step further, use your data to create your audience with the 25% of “purchasers” so you can effectively retarget them through your ad campaigns. When you build your ad campaigns, you’ll retarget this custom audience you built so that you will be only reaching your highest quality audiences.
Too many companies are wasting their time sifting through low-quality leads and even worse– retargeting low quality leads. By getting super strategic with your approach, you’ll ensure you only retarget your best audiences. And this will result in higher conversions overall.
Let’s face it, spreadsheets are still everywhere in Sales and Marketing. We use them for forecasting, we use them for prospecting, we use them for data scrubbing. Yes, the CRM is the source of truth, but sometimes all we need is the speed and flexibility of a spreadsheet
We’ve heard countless stories about our users nightmares about scoring lists in spreadsheets.
Just like a bad dream, the scenario is invariably the same. You receive a prospection file for an event, it contains countless rows of attendees from a wide range of companies, many of which you’ve never heard from. Your goal is to determine if it is worthwhile for your reps to contact these accounts or prospects in order to get an in-person meeting the day of the event. You might also be asked to determine if it is worth sponsoring the event according to the list. Not only is this a critical part of the marketing work, it’s a job-to-be-done, and it’s tedious…
After hearing countless marketers share their struggle to score CSV files we decided to build and launch MadKudu for Google Spreadsheets. This tool allows to directly qualify contacts and accounts in Google Sheets with a simple click.
We’ve been testing MadKudu for Google Sheets in beta for the past few weeks with some of our great customers, and we are really excited to share some of their use cases with you!
One of our clients has a Partnership team whose aim is to gather more customers through their existing partners’ customers. Naturally, it is difficult to get access to the entire clients database. Besides, one of the potential referred companies might seem good for the customer but not for the Partnership team. How can you avoid this situation? The Partnership team has shared their MadKudu API key to the customer, and they run the MadKudu for Google Sheets directly on their spreadsheet. Once the file has been scored, they only share companies with high lead scores. Once a deal is won, a percentage is earned by the customer! It’s a win-win situation :)
Some companies target accounts using specific technologies or tools. Thanks to technographics tool such as Datanyze, you can get a list of all websites using certain technologies. But most of the time, these lists are very long and full of companies outside of your target market. MadKudu for Google Sheets allows them to make sure they will import only the relevant companies in the CRM! True story!
Today’s buyer has evolved. The rise of smartphones, messaging apps, and other groundbreaking technologies has led to a new set of expectations for buyers. They can get exactly what they want in real-time, on-demand, whether that’s scheduling a ride, booking a place to stay or renting a movie.
Of course, these expectations are quickly carrying over to B2B sales teams too. Forms and lengthy follow-ups are out. Live chat and real-time conversations are in.
“A recent study from Twilio showed that 9 out of 10 consumers said they want to be able to use messaging to talk to businesses”
Thousands of businesses are embracing this reality.
More and more companies install live chat like Intercom. Intercom is a leading solution for live chat that’s proven to help you convert visitors with intent.
It becomes a big priority, if not the #1 priority, for a lot of CMOs because it’s a new way to generate SQLs. But handling the live chat is not an easy story with a happy ending.
Live chat is amazing when it comes to sending automated messages. However, automated messages don’t always lead to conversions, and sales teams usually don’t want to talk to unqualified leads/customers/free users.
How many sales come to you complaining about the quality of leads coming to the website? Automated messages allow you to scale your live chat. But does it make sense to be able to automate all the messages you send, yet not be able to qualify faster?
The idea is to unlock modern marketing plays and allow you to schedule qualified calls while you sleep.
At our webinar, we’ll share the top 5 plays we see modern marketing teams using on live chat solutions like Intercom, and we’ll give you a rundown of how we designed MadKudu’s Intercom integration to make it easier than ever to scale live chat.
When SaaS companies first start welcoming customers on board, HubSpot is a natural choice. Its pricing scales with your needs and its tools reach across the entire buyer journey: from lead generation to customer acquisition. By using HubSpot’s various engagement tools like landing pages, ad management, live chat & email campaigns, businesses generate valuable data about what turns prospects into customers.
As they grow, SaaS companies like InVision, Deputy, Front & AppCues complement HubSpot with best-in-breed tools for mission-critical tasks: product analytics data to track user onboarding & activation, data enrichment to learn more about user behavior and profile, and a unified platform for customer data generated by your product and leveraged by all teams.
As data and engagement tools spread, much of HubSpot’s core marketing & sales tools scale very well; however, SaaS companies struggle to leverage all of their customer data inside HubSpot. They want to feed MadKudu’s intelligence & signals into HubSpot to build better buyer journeys.
Up until today, there have been only two ways to score leads in HubSpot: building your own lead score manually using their point-based framework, or using their predictive lead scoring solution. Manual lead scoring requires constant upkeep, tweaking & analysis to make sure you’re identifying the 20% of the leads that will generate 80% of revenue. Manual scoring is most often based on preconceptions, which means it misses non-obvious leads that are actually a great fit in disguise.
HubSpot’s predictive lead scoring solution only leverages HubSpot data and misses out on the rest of your 1st party data and enrichment tools.
That’s why we’re super excited to release MadKudu for HubSpot, baking MadKudu into the entire HubSpot marketing & sales suite of tools. Out of the box, you will be able to:
Use all your customer data inside your scoring (Segment or Amplitude for in-app activity, Stripe for billing…)
Have the lead/account score available in all your other tools so you can optimize the entire funnel
Get account scores, for simple and powerful ABM execution
Understand why each lead is scored the way it is, thanks to our powerful signals and correlations.
We’ve been testing MadKudu for HubSpot in beta for the past few weeks with our great customers. We’re really excited to get it into more HubSpot users’ hands. You can learn more about MadKudu for HubSpot here
Ask a startup how they generate new customers and, chances are, they’ll talk in terms of marketing qualified leads and sales qualified leads.
Marketing qualified leads (or MQLs) are website visitors who have the potential to become great customers: they’ve actively engaged with your company, and they look like a good fit for getting value from your product. Sales qualified leads (or SQLs) have gone a step further, taking an action that suggests a willingness to actually buy from you.
But only 13% of marketing qualified leads ever become sales qualified, according to a B2B sales benchmark report. Those who do require 84 days of education and nurturing to be deemed “sales ready.” Even then, just 6% of those supposedly “sales-ready” prospects ever become paying customers.
Thankfully, there’s another way—a qualification framework that generates active, engaged leads that require less effort to close and make better long-term customers. Instead of MQLs and SQLs, we need to start talking about PQLs—product qualified leads.
What are product qualified leads?
When a lead becomes marketing or sales qualified, we’re using their interactions with our website—downloading an eGuide, viewing a pricing page, or submitting a contact form—to predict when they’re ready to buy.
We might reach out to try to arrange a sales call with somebody who just wanted to read a free whitepaper or cold email a website visitor who clicked onto a pricing page by mistake. This common issue is having someone from the right-fit company being captured as a lead, but not being hot enough to generate a conversation leads to a high volume of MQLs but a low volume of SQLs.
Instead, the PQL—product qualified lead—framework is radically different, using in-app behavior to work out exactly when a lead is ready to purchase.
The PQL model works because SaaS companies can be set up to allow users to get value from the product before making a serious commitment:
Freemium products, like Airtable, offer both free and paid packages.
Free trial products, like ProsperWorks, offer an introductory time- or feature-limited product experience.
Self-service products, like Zuora, allow users to spend as little (or as much) as they like.
This, in turn, allows you to identify when leads become sales-ready with much greater accuracy. Instead of relying on a proxy of sales readiness, it’s possible to see how users actually engage with your product. It works by identifying actions that show users have explored the product and received enough value that they’re ready for a sales conversation.
Product qualified leads are also much easier to turn into customers. The MQL/SQL framework puts the cart before the horse, selling to users before they’ve successfully adopted the product. Sales reps need to do more education and persuasion before a sale, and customer success reps have to work harder to ensure adoption afterwards.
By definition, product qualified leads have already adopted your product. They understand how it works, and they’re getting value from it. As a result, the sales process is little more than a formality. In fact, PQLs close customers at 6x the rate of SQLs, according to Redpoint Ventures partner Tomasz Tunguz.
How to find PQLs
In its simplest form, identifying product qualified leads requires a behavioral trigger: an in-app action or series of actions that correlates most often to users being sales-ready. When a prospect takes that action, they become product qualified and get passed on to the sales team.
For a company like Expensify, that trigger might be the submission of 10 invoices. For Slack, it was a team sending 2,000 messages.
“Based on experience of which companies stuck with us and which didn’t, we decided that any team that has exchanged 2,000 messages in its history has tried Slack—really tried it.” –Stewart Butterfield, Slack
At that 2,000 message milestone, a typical team of 10 people has used Slack for a week. They’ve experienced the product’s core features—quick collaboration, fewer email exchanges, rapid file sharing—and seen firsthand the impact Slack could have on their long-term productivity. Once a team has hit that milestone, they have a 93% retention rate.
Using a tool like MadKudu, these behavioral triggers can be combined with demographic data—looking both at customer fit and likelihood to buy—to fully automate the lead qualification process. Pushing the MadKudu score to sales tools like Salesforce allows sales reps to view a detailed breakdown of every lead’s current qualification status, while allowing Ops teams to prioritize lead routing as well.
Working out manually which actions correlate with sales-readiness can be a tedious process best left to MadKudu; product analytics tools like Amplitude play a valuable role in the modern marketing stack.
More than just a great way to feed precious behavioral data to MadKudu, products like Amplitude help product and marketing teams to dive in granularly into the customer journey—allowing you to drill-down into the last five actions taken by your users—which can help identify opportunities to improve the customer journey and inspiration for experiments.
By tracking conversion rates resulting from each of these actions, you can learn which behavioral signals are indicative of sales intent and adjust your definition of a product qualified lead to match.
How to prioritize PQLs
For a company that generates thousands of leads, it isn’t always enough to have a binary, qualified/unqualified system. You also need a way to prioritize following up with those leads.
This was a problem experienced by HubSpot’s VP of Product, Christopher O’Donnell. When he dug into the company’s acquisition strategy, he identified four distinct sub-types of product qualified leads:
Free users who have hit a given PQL criteria
Users who have requested sales assistance
Users who have reached a limit in their free plans
Users who have purchased without any sales involvement
At one end of the spectrum, we have a free user who has triggered our PQL criteria, completing an action like sending 2,000 messages or submitting 10 invoices. Though this user has experienced the core value of our product, there’s no guarantee they’re ready to buy: they might be happy sticking to our free plan.
Of all our PQLs, these users will require the most sales involvement to become customers: they need to be persuaded that a paid plan is significantly more valuable than the free plan they’re currently using.
Most companies define “Hand raisers” as PQLs who have filled out a contact form to explicitly ask for sales support. They’re more sales-ready than other free users but might still require a few sales touches to convert.
Next are users who have tripped a feature limit on their free plan, like a Slack user reaching their message history limit. These users have successfully adopted the product, so much so that their usage has outstripped their free plan, requiring very little persuasion to become paying customers.
At the far end of the spectrum are “touchless purchases,” free users who are so sold on the value of the product that they become fully-fledged customers without any sales action.
Each of these users is product qualified, but they differ in how much nurturing they need to become customers. By identifying multiple product qualification criteria, it becomes easier to recognize these differences and tailor the sales process to their exact needs.
How to turn PQLs into customers
Instead of trying to persuade someone to purchase a product they’ve never used, you’re helping them get more value from a product they already love. Instead of guessing their needs from a contact form submission or the pages they’ve visited on your website, you can actually see how they’re using your product and shape your sales process to match.
You can find a lead’s most-used product feature and talk through the extra functionality that comes with a paid plan.
You can dig into the limitations they’ve hit on their free package and help them choose a tailor-made plan that matches how they actually use your product.
You can work out which actions they need to take to get more value from the product and send personalized nurturing emails and in-app reminders.
By switching to PQLs, you’re no longer reliant on slick sales patter to close deals. You’re letting your product—the heart of your business, and the reason your customers part with their money each month—sell itself.
Putting customer experience before customer revenue
Product qualification isn’t just a way to speed up sales and improve conversion rates. By using in-app engagement as your primary qualification metric, you’re making a clear, unequivocal declaration: customer experience comes before customer revenue.
Instead of focusing your energy on selling a product to people who have never used it, you’re offering a great product experience, up front, and using the strength of your product to sell itself. You’re using your sales team to support the product and making product adoption your number one priority. You’re generating leads that are easier to close—and most importantly—creating customers who are guaranteed to get value from your product.
Identifying sales-ready accounts to reach out to is the heart of freemium sales acceleration. Freemium businesses rely largely on product adoption to trigger the a-ha moment that will ultimately lead to successful sales engagement, so quantifying that a-ha moment in the form of activity scoring is crucial.
For Sales SaaS reps with hundreds or thousands of accounts assigned to them, reaching out manually every three months yields little or no results. What are the odds that today is the day that an arbitrary account is ready to buy? Very little, right. Coupled with that is the fact that an account may have dozens of associated contacts to choose from and engage. Randomly picking based off of job titles and reaching out is a spray and pray strategy that yields equally unpredictable results.
Sales wants to know which accounts to reach out and who within that account to reach out. And we’ve got just the play for that.
Timing is everything: Surfacing sales-ready accounts & the right contacts to engage.
Our goal here is for sales reps to start every day with a list of accounts assigned to them that are ready to have a conversation. We’d also like to give that sales rep a filtered list of the contacts most likely to respond. That way sales reps spend all of their time crafting the most relevant message for their best leads.
This is a great play for freemium SaaS businesses with a large number of accounts where velocity & efficiency are key to success. This play is also good for products with a combination of low-touch and high-touch users – while you may have paying customers already, identifying when that paying account is sales-ready is key.
The bulk of this play is going to sit on MadKudu’s ability to build an accurate account-level behavioral scoring model. Structured data is going to be key – wecan’t build an account-level model if we don’t have account data. When we run this play with InVision, we use Segment for product data & HubSpot for marketing data. We’ll also need Salesforce for our sales data.
Once we have our data piping correctly, MadKudu is going to identify the features & activity that best predict sales readiness by looking at historical product & marketing engagement and the resulting sales outcome. Once the predictive model is build, MadKudu sucks down the latest activity data on a regular basis and looks for triggers.
When a sales-ready account is identified, MadKudu tags it in SalesForce and drops it into a daily report for sales reps. Within each sales-ready account, MadKudu’s models also look at the profile & activity of each contact in order to identify the contacts most likely to engage. Those contacts are recommended to the sales rep.
The Impact: +25% in Pipeline
InVision’s 25% increase in pipeline came from identifying accounts programmatically based on historical customer data. This came without any change to the product, just in optimizing for sales-readiness. If you’re looking at lead activity today, you may be leaving money on the table. With InVision, we identified accounts where no single lead achieved the activity threshold for an MQL, but the account as a whole hit the MQA threshold. It is the combined activity of several users that makes an account ready to talk to sales.
Once this MQA model is built, we can begin to add some of our other plays on top of it. Forcing the funnel by sales-ready accounts by customizing the app or website. Reducing friction on forms or triggering chat for enterprise prospects with a Fastlane play. We can score accounts throughout the buyer journey.
90% of your website traffic doesn’t convert, and there’s nothing worse than a missed opportunity. For b2b companies, retargeting is a no-brainer. It’s an easy way to make sure you’re always targeting an audience that has showed some intent to buy. However, the problem with retargeting anonymous website traffic broadly is that you don’t know who you are targeting and how qualified they are.
With just a high volume, many SaaS companies bid low on retargeting across their entire website traffic. They push their brand in front of their website traffic wherever they go. This spray and pray tactic means SaaS companies are only getting in front of traffic that other advertisers aren’t willing to pay more. Do you think your competitors may have a more focused strategy, outbidding you on your best leads and leaving the rest to you?
Click through rate is low because the quality filter is low. Conversion rate is low because most of your website traffic shouldn’t convert (candidates, low-quality leads, investors, analysts, perusers).
Our goal here is to focus our retargeting budget on the subsection of our website traffic that is worth the most to us. If we do that, we will be able to reallocate the budget we’re not spending on low-quality traffic to bidding more for our high-quality traffic.
We’ll need a few tools to Re-Get That Bread:
IP lookup: we’ll be using Clearbit Reveal for this.
Qualification: we’ll be using MadKudu for this.
Retargeting: we’ll be using Adroll for this.
As usual, we’ll be connecting this all through Segment.
Qualifying traffic has become pretty easy with the advent of IP Lookup APIs – the most popular being Clearbit Reveal. Feed Clearbit an IP address and it returns (among other things) the domain of the company or of the individual visiting your website. This is enough to score an account. We’ll be scoring with MadKudu, but you can also do it with your homegrown Lamb or Duck lead scoring model. We’ll send MadKudu the domain name provided by Clearbit, which will return a few important data points for this play:
Customer Fit segment: very good, good, medium, low
Predicted Spend: custom based on your specific pricing plans, predicting which
Topical Segmentation: custom based on your target segments (e.g for Algolia: ecommerce, media, SaaS).
With this data we’re able to feed AdRoll a custom audience of qualified traffic to target. This can be a bit tricky since AdRoll requires a static audience, but a quick script to update a static audience on a daily basis will get us around that hiccough.
Based on predicted spend, we can even build separate audiences for our various plans, each with different budgets. If we add in Topical Segmentation, we can run targeted messaging to our various ICPs based on their needs at various price points. If we know the predicted value of the qualified traffic, we can calculate our maximum budget as a function of our acceptable CAC.
The Impact: +300% click-through rate
When Chris Rodriguez at Gliffy first began building this play, he was looking to get click through rate for his retargeting ads under control. When he saw it jump from the .7% industry average to 2-3% for qualified traffic, it became pretty clear that qualified traffic was worth the focus.
Bidding higher on a qualified audience is a no-brainer: we see it on ad networks that boast a qualified audience or a qualified system of manual segmentation. It only makes sense that we would apply the same logic to how we retarget our own audience: we want to spend more on the audience that matters, the ones that got away.
Ten years ago Amazon introduced same-day delivery, probably the single most important feature in cementing their dominance of the eCommerce industry. They did this after 10 years of innovating on the online shopper experience – recommended purchases, one-click payments, experiments on how website latency affected conversion rates – and they understood that the biggest source of friction in their buyer experience was waiting for your package to arrive.
We all have an idea in our head about what makes a great customer journey, a great buyer experience. When Francis asked me out of the blue, the first thing that came to mind was my experience buying an engagement ring last year, but I could just as easily point to the experience of creating a new Slack team. They are magical experiences. You never see what’s going on behind the curtain, and you never have any downtime to think about it. Is my package already in Paris? How did Amazon know what I was going to order? Doesn’t matter. It’s already arrived before I can begin to comprehend how they possibly do that at scale.
For SaaS companies today, increasing revenue is often about removing friction. The product team designs and improves features so that customers don’t have time to wonder whether the competition is building a better product. Customer Success is looking at customer health metrics to identify customers before they even think about churning and improve their results.
Marketing & Sales have a plethora of data & engagement tools so that they know everything about who their engaging with from Clearbit-enhanced Drift Bots to segmented Outreach campaigns encouraging prospects to signup for Webinars or jump on a call.
You are the friction.
So why is it that 90% of SaaS companies take more than five minutes to follow up on a request to schedule a demo? Francis suggests going through your own customer journey – ideally by signing up with a friend’s email account, especially if your friend is a great fit for your product – to get the full experience. If it’s not the ~48 hours of follow-up time that’ll make you feel the friction, it’s the ~5 days between the demo request and the phone call that’ll make you rethink your process.
What makes it take so long?
Lead data enhancement
SDR first-response latency
Email back-and-forth to validate interest and find a time to talk.
It’s easy to understand each one of those steps – after all, everything above (accept maybe the emails) feel very logical – the only thing that’s missing in the equation is the customer experience. SaaS companies are eager to over-optimize for the sake of being fair, applying rigorous rules to lead assignment, and this often flies in the face of the customer journey.
One of MadKudu’s most popular features, the Fastlane – an enhancement to signup forms that allows highly-qualified leads to skip the form and go straight to a sales rep’s calendar – is often difficult to implement initially because lead routing takes minutes. The customer eats the friction because of operational friction.
Remove friction. Prioritize customers.
It’s easy to remove friction from the customer journey if you prioritize it. Calendly, for example, offers a great Team Scheduling feature that allows prospects to see an aggregate calendar for every potential representative and then choose a time that works for them, instead of displaying the representative’s calendar after they’ve been round-robined with less available time slots. This puts the customer in the priority seat and sacrifices the possibility that Rep’s who have less immediate availability in their calendar might get routed less leads. In fact, that’s not a bad forcing function for making sure SDRs are prioritizing their time correctly.
Is the AI hype starting to wither? I believe so, yes.
Reality of the operational world is slowly but steadily catching up with the idyllic marketing fantasy. The report Jeffries put together challenging IBM Watson proves alarm bells are ringing. The debacle of the Anderson implementation goes to show how marketing promises can be unrealistic and the downfall will be dreadful.
With that said, not all is lost as we keep learning from our past mistakes. Being part of the Salesforce Einstein focused incubator, we are witnessing first-hand how the CRM giant is looking to succeed where Watson and others are struggling. Hopefully these insights can help others rethink their go-to-market strategy, in an era of unkept commitments.
Salesforce, a quick refresher
A few weeks ago, I was being interviewed for an internal Salesforce video. The question was “how has the Salesforce eco-system helped your startup?”. To contextualize my thoughts it’s important to know that while Salesforce is one of our main integrations, we consider it as an execution platform among others (Segment, Marketo, Intercom, Eloqua…). I’ve always admired Salesforce for its “platform” business model. Being part of the Salesforce ecosystem facilitated our GTM. It gave MadKudu access to a large pool of educated prospects.
However I believe the major value add for startups is the focus induced by working with Salesforce customers. Since Salesforce is a great execution platform there are a plethora of applications available addressing specific needs. This means, as a startup, you can focus on a clearly defined and well delimited value proposition. You can rely on other solutions to solve for peripheral needs. As David Cohen reminded us during our first week at Techstars, “startups don’t starve, they drown”. Salesforce has helped us stay afloat and navigate its large customer base.
What is Salesforce Einstein?
I’m personally very excited about Salesforce Einstein. For the past 5 years, I’ve seen Machine Learning be further commoditized by products such as Microsoft Azure, Prediction.io… We’ve had many investors ask us what our moat was given this rapid democratization of ML capabilities and our answer has been the same all along. In B2B Sales/Marketing software, pure Machine Learning should not be considered a competitive advantage mainly because there are too few data sets available that require non-generic algorithms. The true moat doesn’t reside in the algorithms but rather in all the aspects surrounding them: feature generation, technical operationalization, prediction serving, latency optimization, business operationalization… The last one being the hardest yet the most valuable (hence the one we are tackling at MadKudu…).
Salesforce Einstein is the incarnation that innovation will be in those areas since anyone can now run ML models with their CRM.
We’ve been here before
Just a reminder, this is not a new thing. We’ve been through this not so long ago.
Remember the days when “Big Data” was still making most of the headlines on Techcrunch? Oh how those were simpler times…
Big Data vs Artificial Intellligence search trends over the past 5 years
There were some major misconceptions as to what truly defined Big Data especially within the context of the Enterprise. The media primarily focused on our favorite behemoths: google, facebook, twitter and their scalling troubles. Big data became synonymous for Petabytes and unfathomly large volumes of data more generally. However scholars defined a classification that qualified data as “big” for 3 reasons:
– volume: massive amounts of data that required distributed systems from storage to processing
– velocity: quickly changing data sets such as product browsing. This meant offline/batch processing needed an alternative
– variety: data originating from disparate sources meant complex ERDs had to be maintained
In the Enterprise, volume was rarely the primary struggle. Velocity posed a few issues to large retailers and companies like RichRelevance nailed the execution of their solution. But the main and most challenging data issue faced was with the variety of data.
What will make Salesforce Einstein succeed
Einstein will enable startups to provide value to the Enterprise by focusing on the challenges of:
– feeding the right data to the platform
– defining a business playbook of ways to generate $$ out of model predictions
We’ll keep the second point for a later blog post but to illustrate the first point with DATA, I put together an experiment. I took one of our customers’ dataset of leads and opportunities.
The goal was to evaluate different ways of building a lead scoring model. The objective was to identify patterns within the leads that indicated a high likelihood of it converting to an opportunity. This is a B2B SaaS company selling to other B2B companies with a $30k ACV.
I ran an out-of-the-box logistic regression on top of the usual suspects: company size, industry, geography and alexa rank. For good measure we had a fancy tech count feature which looked at the amount of technologies that could be found on the lead’s website. With about 500 opportunities to work on, there was a clear worry about overfitting with more features. This is especially true since we had to dummy the categorical variables.
Here’s how the regression performed on the training data (70% of the dataset) vs the test dataset (30% that were not used for training and ensuring if a company is part of the training it is not part of testing – see we did not fool around with this test)
model performance on test dataset using available data points
Not bad right?! There is a clear overfitting issue but the performance is not dreadful apart for a blob in the center
Now we ran the same logistic regression against 2 feature: predicted number of tracked users (which we know to be highly tied to the value of the product) and predicted revenue. These features are the result of predictive models that we run against a much larger data set and take into account firmographics (Alexa rank, business model, company size, market segment, industry…) along with technographics (types of technologies used, number of enterprise technologies…) and custom data points. Here’s how the regression performed:
model performance on test dataset using 2 MadKudu features
Quite impressive to see how much better the model performs with fewer features. At the same time, we are less running the risk of overfitting as you can see.
The TL;DR is that no amount of algorithmic brute force applied to these B2B data sets will ever make up for appropriate data preparation.
In essence, Salesforce is outsourcing the data science part of building AI driven sales models to startups who will specialize in verticals and/or use-cases. MadKudu is a perfect illustration of this trend. The expert knowledge we’ve accumulated by working with hundreds of B2B SaaS companies is what has enabled us to define these smart features that make lead scoring implementations successful.
So there you have it, MadKudu needs Salesforce to focus on its core value and Salesforce needs MadKudu to make its customers and therefore Einstein successful. That’s the beauty of a platform business model.
I also strongly believe that in the near future there will be a strong need for a “Training dataset” marketplace. As more of the platforms make ML/AI functionalities available, being able to train them out-of-the-box will become an important problem to solve. These “training datasets” will contain a lot of expert knowledge and be the results of heavy data lifting.
MadKudu demo Jam 3
PS: To be perfectly clear, we are not dissing on IBM’s technology which is state of the art. We are arguing that out-of-the-box AI have been overhyped in the Enterprise and that project implementation costs have been underestimated due to a lack of transparence on the complexity of configuring such platforms.