Startup vs bigco: the best career option in data science

Our fearless leader and CEO Sam Levan recently spoke at Galvanize in San Francisco about data science careers. A common student question is, “is a job at a startup or a big company better for data scientists”?

At MadKudu we’re in a good position to answer it – the 5 of us have more than 25 years experience in data science. We’ve built everything from the world’s largest fraud detection system to quick hacks in Google sheets.

So what is the best career option for an aspiring data scientist? Google or HotNewStartupWithCuteBlueDoggyMascot.com?

Before answering I’m going to share a secret about being a data scientist. Whether you work at a 2-person startup or CapitalOne, there is one attribute which best predicts your probability of success.

The super-duper-secret to being a great data scientist is …

… wait for it …
…… wait for it ……
……… WAIT FOR IT!!! ………

Data science is a SOCIAL skill

That’s it. That’s the big secret. Your career as a data scientist will be defined by how well you can communicate, write, listen, organize, lead, and empathize.

Do you need hard skills? Of course. You can’t do the job if you don’t know how to use R, Python, or MatLab. You have to know how to measure the statistical significance of your results.

But unless you fancy yourself the next Will Hunting, being a brilliant Python coder won’t make you any more effective than a good one if you can’t work well with others.

Data scientists don’t work alone

Data scientists try to solve business problems with data – an iterative activity which requires working with cross-functional teams.

Suppose you’re a knowledge engineer working at a bank. On any given day you will need to:

  • Talk to regulators about the riskiest type of criminal activity.
  • Help analysts understand customer data and what behavior the bank can track.
  • Ask (beg?) the operations to pull new data sources for you.
  • Testify in court.

The most effective data scientists are team players who make everyone else more effective.

If you want to be a lone hero data science isn’t for you.

Communication beats code

Every morning I read Nate Silver’s analysis on FiveThirtyEight. Is Nate the world’s greatest statistician?

Of course not. Nate Silver’s brilliance is his ability to help us understand how data answers important questions in politics, sports and life.

We try to deliver the same value in our work.

This is – by far – our most popular blog post. Why was it featured on Growth Hackers, Hacker News, and Growth Hacking digest? It wasn’t a great study – just 9 sample companies. We didn’t build any amazing models – everything was done in Google Sheets.

As an example of data science it is … meh. Our customers loved it because we helped them understand what the data means and how they can use it to solve business problems.

Any idiot can have an opinion. Lots of smart people can compile numbers.

Few people can help others understand why data matters and what they should do – be one of them and you’ve got the world at your feet.

Data science careers: Startups vs Bigco

Back to the students’ questions:

“What is the difference between being a data scientist at a startup vs a Bigco?”

Specialist vs generalist

The biggest difference between being a data scientist at bigco vs a startup is your degree of specialization. At a bigco you have the opportunity to work for months and years on the same problem.

Are you excited about spending 2 years creating the world’s greatest recommendation engine for Facebook? Do you like doing primary research? Becoming an expert in building models to solve 1 problem? Becoming a master in R, Python, or MatLab?

That’s life in a bigco. I know Knowledge Engineers who spend a career detecting violations of stock market wash sale rules.

Startups? Ha ha ha!

You won’t know what you’re working on next week, much less next year. You’ll be spending your time helping marketing, sales, and product teams answer basic questions. Since you’re constrained by data and time you’ll do much of your work in spreadsheets or SQL.

It isn’t uncommon for a data scientist at a startup to be juggling 5 different problems at the same time. Your expertise will be your ability to quickly acquire and apply new skills – fortunately this is a great skill to have.

Support system

Unless you work at MadKudu, you may be the only data scientist at your startup. Your colleagues may not understand what you do or how you can help them. You may have to define your own objectives. On your first day you might be told to “go help the sales team find the best leads”. Does this terrify or excite you?

At bigco you will have a support system. Your boss will tell you which project you’re working on. More experienced data scientists can answer your questions. Have a problem? Ask your boss – that’s what she’s for.

Getting dirty

Data science textbook examples are fairy tales. In 20 years I’ve never encountered such simple problems. In the real world:

  • Simply getting the data is HARD.
  • People don’t agree on what columns actually mean.
  • Everything changes while you’re doing analysis.

Bigcos have teams of people to help solve these problems: server-side developers to populate the data warehouse and business analysts who write data dictionaries.

At startups … well … it is probably up to you. The developers are all too busy finishing the next release and supporting customers to run SQL queries. You have look in the code to see how the product generates the account_activated event in Mixpanel.

Risk

At bigco you’ll have a nice salary, 401(k) plan, and benefits. You’ll work a little harder the 3 months before bonus time so you can get that new car. It feels safe – but is it?

Life at a startup is the opposite. Part of your compensation will potential, unknown upside from stock options. Will you have a job next quarter? It depends on whether the CEO can close the B round. It feels risky – but is it?

I’ve worked for the world’s biggest, most stable employer and at any-day-we’re-dead-lets-start-stealing-office-supplies startups. I’ve had friends lose $500K starting a company and others struggle for years to find a job after being laid off. Here is how I think about risk.

Working for startups is very risky in the short run but incredibly stable in the long run.

The stability of bigcos comes at a price – you develop fewer skills, build fewer relationships, and don’t get regular experience marketing yourself.

This risk is particularly true for a data scientist who can get stuck working on the same problem … with the same tools… and the same people … for years. A major industry downturn can be economically devastating when all companies in a sector are laying off employees.

Both bigco and startup careers have risks – you just need to understand the risks you’re taking and be smart about managing them.

What’s best for you – bigco or a startup?

After reading this post you’re probably more confused than ever – because there is no one answer.

My #1 piece of advice is to go out an interview with big and large companies. Meet the teams and ask lots of questions.

What is the #1 problem you would be solving? Why? Who else is on your team? What do they say about the problem? What tools would you be using?

It’s the only way you’re going to see what is best for you.

Best of all it will give you an opportunity to work on those social and communication skills – the most critical ones for your career in data science.

Photo credit: goingfar.org

How to accelerate SaaS sales with behavior-based conversions (part 2/3)

This is Part 2 of a 3-part series on how to build a Behavior-based conversion strategy for your SaaS product. Here is Part 1.

This post was a lot of fun to write – I hope you enjoy learning from the insights as much as I did.

In a previous post, we analyzed conversion data from 9 SaaS companies and concluded that optimizing conversions based on behavior is more effective than using an X-day trial for every customer.

IOW, starting the same 30-day trial for every new customer isn’t as effective as varying the buying incentives based on what the customer has done – what we call a behavior-based conversion strategy.

3 steps to begin a SaaS Behavior-based conversion strategy

30-day trials are a great way to start a SaaS company. They are easy to setup and the trial expiration creates a “buy now” urgency. But using the same 30-day trial for every customer isn’t as effective as putting

… the right purchasing incentives to
…… the right customers at
……… the right time.

In Part 1 I described the 3 small changes required to get started.

  1. Don’t begin a timed trial until a customer completes Activation events – those a trial customer must complete or they won’t convert.
  2. During the trial use marketing and sales to get customers to complete Engagement events – the key value-creating used by most of your customers.
  3. Create early conversion and upsell incentives for customers who complete Acceleration events – “delight” events which indicate customers are getting tremendous value from your product.

Activation events are “NECESSARY but NOT SUFFICIENT” – customers who don’t complete them don’t buy but completing them doesn’t lead to a conversion.

Engagement events are “NECESSARY and SUFFICIENT” – customers who don’t complete them don’t buy and completing them increasing conversions.

Acceleration events are “SUFFICIENT but NOT NECESSARY” – customers who don’t complete them might buy anyway and completing increases conversions.

table

In this post I’ll cover how you can find Activation, Engagement, and Acceleration events and build your Behavior-based conversion strategy around them.

Use the Behavior conversion matrix to find your Activation, Engagement, and Acceleration events

The Behavior conversion matrix will help you instantly identify these events based on how your customers are actually behaving.

First let’s use your data to categorize events based the theoretical “Necessary” and “Sufficient”.

Categorizing “Necessary” and “Sufficient” events in your data

Identify ~50 relevant boolean customer events and calculate the following:

  1. The percentage of time the customer DID the behavior and converted to a paying customer. This is the probability of conversion if TRUE – the degree to which an event is “Sufficient”.
  2. The percentage of time the customer DID NOT do the behavior and converted to a paying customer. This is the probability of conversion if FALSE – the degree to which an event is “Necessary”.

(I’ll show you step-by-step how to calculate these in Part 3 of this series)

Here is a summary of how these conditions relate:

summary table

Create the Behavior conversion matrix by calculating event probabilities

By taking every customer event and calculating these probabilities we can now identify which ones are Activation, Engagement, and Acceleration events. This is most easily illustrated with an example.

Imagine a SaaS small business accounting application. Most customers use it for basic expenses tracking and reporting. The most valuable customers also find and hire a tax accountant through the application.

We can calculate the conversion probabilities for each event as follows:

tables

Which can be plotted in a bubble plot as follows:

On Average, 3.1% of all trial customers convert to paying customers. The Behavior conversion matrix allows us to see how different events relate to the Average.

signed_up is – not surprisingly – the simplest example of an Activation event. A customer who doesn’t sign up can’t possibly pay, but since every customer signs up the event doesn’t improve our chances of converting. Thus the probability of converting if signed_up == FALSE is 0, while probability of converting if signed_up == TRUE is 3.1% – the same as the Average user.

connected_bank is a more interesting Activation event. Very few customers who don’t connect their bank to the application convert – perhaps they love entering data by hand for some sadistic reason – but taking this step doesn’t improve their conversion rates.

entered_expense is an Engagement event. A customer who doesn’t complete has <1% chance of converting while those who do are 4x likely to convert (12% / 3.1%).

hired_accountant is an Acceleration event. Customers who find an accountant through the app will almost definitely buy, but those who don’t still often will.

Accelerating SaaS Sales with the Behavior Conversion Matrix

Identifying your most important conversion

After plotting dozens of events your Behavior conversion matrix will start to look like this:

graph2

The most important events are those distant from the Average User Baseline – you will get the most impact by organizing your onboarding around them.

In 3 Steps – from 30-day trial to Behavior-based conversions

Suppose our simple SaaS accounting app currently starts all customers on a 30-day trial at signup. Here are a few simple changes to implement Behavior-based conversions.

  1. Begin the trial after the customer completes Activation events

Continue to nurture a trial customer until connected_bank is TRUE and then start the 30-day trial – until he takes that step he isn’t in a position to buy. He doesn’t know the value of the product and a trial expiration won’t create a “buy now” incentive. Connecting a bank account to the app is a Necessary but NOT Sufficient event for conversions.

  1. Create incentives to complete Engagement events during the trial

During the trial use marketing automation and inside sales to get the customer to complete entered_expense. Offer workshops on cost accounting or have Customer Success reps help them choose expense categories. Offer trial extensions and continue to convert these customers for up to 6 months following trial expiration. Entering an expense is a Necessary AND Sufficient event for conversions.

  1. Create early conversion and upsell incentives for customers who complete Acceleration events

Upsell customers to an annual plan or offer early-conversion discounts after the event hired_accountant. These are the most promising prospects and many will convert early. Hiring an accountant is NOT Necessary but is a Sufficient event for conversions.

In Part 3 of this series I’ll walk through a complete end-to-end example showing how derive the probabilities above. We plan on releasing it in mid-April – if you would like to see it sooner please tell us on Twitter or comment below.

Like our writing? You’ll love our product. Try MadKudu for free.

 

Photo credit: Timothy Neesam

How to accelerate SaaS sales with behavior-based conversions (part 1/3)

In a previous post, we analyzed conversion data from 9 SaaS companies and concluded that optimizing conversions based on behavior is more effective than using an X-day trial for every customer.

We call this approach “behavior-based conversions”. In this 3-part series I’ll explain the strategy and show you how to implement it.

Subscribe to our newsletter I’ll send each to you as I write them.

The pros & cons of your free 30-day trial

Easy to setup and manage

X-day free trials are effective because they reduce the cost of sales and create an artificial purchasing incentive. Customers try your products risk-free using their own data and you have the opportunity to delight them.

The trial deadline creates an artificial “buy now” incentive that drives conversions.

X-day trials are easy to create and manage with payment APIs like Stripe and Recurly. Just start the trial when customers sign up and try to get them to pay before it ends.

But not optimal if you treat every customer the same

X-day trials are not optimal because customers will convert at different rates. An expiring trial isn’t the best purchasing incentive for all of them. Some would pay right away, some need a few days, and some take 6 months.

Looking for the “optimal” trial length is a fool’s errand – there isn’t anything magical about a 14-day, 30-day, π-day, or 65.5890987899875476-day trial.

Your customers are all different.

Behavior-based conversions accelerate sales by treating customers differently

A customer may need 2 months of nurturing before buying a product – if so, giving him only 14 days risks losing a sale. Contrarily, if a customer is ready to buy at signup we lose 30 days of revenue by asking him to pay at the end of a trial.

Behavior-based conversions get

… the right purchasing incentives to
…… the right customers at
……… the right time.

Creating your behavior-based conversion strategy

Don’t begin timed trials until a customer is able to make a buying decision

Timed trials are great for creating artificial purchasing urgency – otherwise customers don’t have an incentive to “buy now”. But they only work if customers understand the value they get from your product.

Every customer needs to take a few early, critical actions before she can even begin using the product – signing up, connecting her Gmail account, etc. Before taking these actions she isn’t in a position to buy because she doesn’t understand the value your product would create for her.

These are Activation events – the critical steps that must happen before your customer begins to see the value of your product. Customers who don’t complete Activation events rarely buy – more importantly the deadline of your timed-trial won’t motivate them to complete Activation events. (this is easily provable as you’ll soon see).

Nurture your customers through marketing automation and sales until they complete Activation events – THEN start your timed trials.

Activation events are “NECESSARY but NOT SUFFICIENT” – customers who don’t complete them don’t buy but completing them alone doesn’t creating enough value for customers.

During the trial focus on creating initial value

Your product performs a few key value-added activities for customers. These are usually the core features of your product, those most of your paying customers use. You can identify key events in the customer lifecycle that indicate a customer is getting basic value from your product – these are Engagement events, or what Lincoln Murphy calls First Value Delivered (FVD).

Engagement events are NOT “delight” events – those associated with the few customers who are getting the most possible value from your product.

Try to get customers to complete Engagement events during your timed trial and use the trial deadline to create an artificial purchasing incentive.

Continue trying to get customers to complete Engagement events past the trial deadline. Offer trial extensions, workshops, sales calls, etc. for up to a year. Don’t simply toss these customers back into the same “nurture” bucket as those who haven’t completed Activation events.

Engagement events are “NECESSARY and SUFFICIENT” – customers who don’t complete them won’t buy but those who do are getting value from your product.

Accelerate conversions for delighted customers

Some customers instantly understand your product and quickly get tremendous value from it. These customers are ready to buy now and you can accelerate your SaaS sales by trying to get them to convert before a trial ends or upsell them.

You can identify these customers by those who complete Acceleration events – activities associated with your most delighted customers.

Offer additional purchasing incentives to customers who complete Acceleration events to get them to convert early.

Examples of purchasing incentives are coupons, annual contracts, or special offers. Or simply tell them to enter a credit card now so they don’t lose service.

Acceleration events are “SUFFICIENT but NOT NECESSARY” – customers who complete them are getting tremendous value but those who don’t might buy anyway.

So … are any real SaaS companies doing this?

Absolutely. In fact, you’re probably already doing some behavior-based conversions by extending trials or using your inside sales teams to close customers faster.

You will find every one of these tactics used by SaaS companies – we just put them together into a unified strategy we call behavior-based conversions.

Summary: Activation, Engagement, and Acceleration events

Here is a summary of the 3 events and how to use them.

table

How to find your events

You’re now wondering, “how do I identify Activation, Engagement, and Acceleration events?”

Your sales funnel diagram won’t help – it only shows what you think your customers should do, not what they actually are doing.

In Part 2 of this post I’ll show you how to identify Activation, Engagement, and Acceleration events by looking at your data.

You don’t need to be a data scientist. You won’t have to learn what “entropy” is and how it differs from a canopy.

You just need Google Sheets, 3rd grade math, and an open mind.

Part 2 of this post

Now the real fun begins – read part 2 of this series.

 

Photo credit: Mark Freeth

Be a hero to your sales team with this Slack hack

Let’s face it – marketing and sales operations can be thankless work. When we generate quality leads we’re “doing our jobs”. We toil away making slow, methodical progress – progress often unseen by the rest of the company.

Would you like to be the hero for a change?  To make a, quick, high-visibility impact with your sales team?

Of course you do! If so, give this hack a try.

With a bit of coding you can create a “hot” lead notification using Clearbit and Slack.

It’s a fun project for a Friday afternoon or a hackathon – we know from experience that sales teams love it.

Use Clearbit’s sweet API to learn more about your leads

What does your sales team do when gophillyeagles998@hotmail.com signs up for a 30-day trial?

Nothing! Your lead-scoring system ignores hotmail accounts.

Now suppose gophillyeagles998@hotmail.com is CEO of a 100-person company in Chicago (where he is the only Eagles fan) and he’s using a personal email account to test our your app.

ooooooops! Your sales team just missed an opportunity to engage a white-hot lead.

Clearbit helps solve this problem. Send Clearbit’s API an email address and it returns information about a lead such as:

  • Name
  • Location
  • Title
  • Company size
  • Industry

With a bit of hacking you can use the API to build a lead notification system in Slack.

Why you’ll be a hero for the VP of Sales

Slack “interrupts” sales reps to call immediately

Most sales teams know an ideal time to call a qualified lead – usually within minutes of signup or after a specific event. Since we all live in Slack they can get an instant notification to take action.

Miss fewer good leads who use a free email account

Some qualified leads will test your product with a personal email account. Identify and give them some extra love to make sure they have a good experience.

Qualify leads based on company size, role, etc.

Sometimes you only have an email address. Sure, you can manually look up businesses based on email domains or ask for additional information but this is a hassle.

With Clearbit’s API you can quickly qualify leads on simple metrics like “send me a Slack notification if company size is greater than 30”.

Add “conversation starters” context

Good sales reps look for anything relevant to get closer to a lead or start a conversation. With a Slack notification it is all there at their fingertips.

“You’re an Eagles fan too? Man, sure glad Chip Kelly is gone. Must be lonely there in Chicago”.

“Since you’re in manufacturing I’m guess you signed up to take advantage of our partnerships in Asia, correct?”

How it works

Send Clearbit an email address from your app (or a separate code hack), determine if the lead is qualified and post to Slack

slack_notify

How to setup

Some custom dev work – but don’t be afraid

Ok, ok, I know. Developer time is the scarcest resource at your company and they’re already overworked. This hack does take some custom dev work but:

  • The Slack and Clearbit APIs are well-documented and written by developers for developers – speaking as one myself, we like working with these type of tools.
  • Even junior level server-side developers can do it. It makes for a fun Friday project or hackathon.
  • You can give developers some very specific requirements about what you want by following the steps below – this will save them a ton of time.

If all else fails just find a bored developer and learn her favorite Starbucks drink – you’d be surprised what you can get done with a nice word and a $5 Latte.

Step 1 – Sign up for a Clearbit account.

They have a free version with limited API calls. The paid plans will pay for itself if you get 1-2 new deals/month from doing this.

Step 2 – Document which fields you want from Clearbit

Login and visit https://dashboard.clearbit.com/docs#enrichment-api and identify what Person and Company information you want in Slack.

Start with just a few of the most important ones – and understand that Clearbit usually only has a subset of this data.

clearbit api

Step 3 – Identify your threshold for qualified leads

Write down the qualification rules for leads.

e.g. “only create a Slack notification if metrics.employees > 50

Step 4 – Create a dedicated Slack Channel for the notifications

slack channel

Step 5 – Send this to your developer

Give your developer the Clearbit login credentials and the requirements you created in Step 2 and Step 3.

(And we’ll be oh-so flattered if you send her the URL to this blog post as well.)

Developer resources

In addition to the Clearbit and Slack APIs there is clearbit-slack on Github. You can also ask us any questions or suggest other resources in the comments below.

Bonus: User predictive analytics to include in-app behavior

This a simple approach for quickly alerting sales teams about qualified leads through Slack. It is a great way to get started if your sales team isn’t identifying and contacting leads quickly.

Add your customer’s in-app behavior to qualify leads better and arm sales with more customer data

Adding predictive analytics based on user behavior is far more effective – especially if you get >10 leads/day. We can do this for you. The benefits of predictive analytics are:

  • Better qualification based on what your best customers have done in the past.
  • More effective “call now” notifications the moment customers are ready to buy.

Predictive analytics does this more effectively than building and managing a pile of lead scoring rules. Best of all we can do this for you so you don’t have to hire a data scientist. And we can do this for you.

 

Did I mention we can do this for you?

30-day trial? 14-day? Freemium? Here’s why it probably doesn’t matter

2/29/2016 update – We’ve had a number of requests to expand on this post and provide examples of behavior-based conversion incentives. We decided to write a 3-part series on this topic. You can read the first one here

Whenever I launch a new SaaS product I obsess about sales and onboarding details.

Should I offer a free trial? How long?
Or should I have a free version with no trial (freemium)?

The blogs and books have opinions but most are based on limited data or anecdotes from one SaaS marketing team. Here at MadKudu we try to answer these questions based on data – how our customers are selling.

This week Erik, Sam and I tried to learn how trial length affects SaaS conversions and revenue acceleration.

Here’s what we learned.

Data from 9 representative SaaS companies

We selected 9 companies with different models and relatively clean data. We then identified every trial user who converted to a paying customer and grouped them by of days it takes to convert.1

Example

Here is a simple example you can copy to illustrate the process.

Screen Shot 2016-02-12 at 4.36.16 PM

2 days after customers sign up for this fictitious SaaS company 142 (87+55) converted, 142/305 = 47% of the total who will eventually convert.

Results

After looking at the number of daily conversions we graphed their accumulation relative to starting a trial and came up with the following.

Screen Shot 2016-02-12 at 4.47.23 PM

How to read this graph

This graph shows the rate at which a SaaS company converts trial customers to sales. For example, here is how evaluate the results for Company C:

Screen Shot 2016-02-12 at 4.57.56 PM

Company C offers a 30-day free trial – not surprisingly, most customers who decide to pay do so at the end of their free trial. But customers also convert before and after the 30th day of the trial. This graph illustrates the rate at which that happens.2

Sales acceleration

You accelerate sales (and get a high-five from Tomasz Tunguz) by pushing this curve up and to the left without sacrificing top-line revenue.

Screen Shot 2016-02-12 at 5.00.04 PM

SaaS companies accelerate sales to hit profitability faster. VCs like Redpoint’s Tunguz look for companies who can get customers to start paying sooner.

Observations

We studied relationships between trial lengths, conversion rates and models – the results surprised us.3

Screen Shot 2016-02-12 at 5.19.57 PM

Observation 1: It takes about 40 days to get 80% of SaaS conversions

It takes most SaaS customers a little more than a month to test out and purchase a product. This general rule seems to hold true regardless of trial length or whether a product has a free version.

This is … well … surprising.

For instance, look what happens when we isolated a SaaS company Freemium (G), 14-day trial (E), and a 30-day trial (C).

Screen Shot 2016-02-12 at 5.29.57 PM

Not surprisingly, freemium conversions are faster since customers can quickly choose to purchase the premium versions. And, of course, a 14-day trial accelerates faster than a 30-day trial.

But these curves converge when 80% of customers convert, around 40 days after a trial starts.

Observation 2: Half of SaaS conversions happen AFTER the trial ends.

A free trial creates artificial purchasing urgency. But there isn’t anything magical about the last day of a trial – some customers continue to convert at their own rate based on incentives or their perceptions of value.

Every single company we studied had customers who converted more than 100 days after signing up – most had customers who converted 6 months after signing up.

Observation 3: The “S” curve is the $ curve

Ok – this is cool. When we first noticed these results we assumed it was an error. It isn’t. Check out how similar the curves are between company B and C:

Screen Shot 2016-02-12 at 5.48.12 PM

 

Identical curve … different businesses.

Why? Both companies have 30-day trial SaaS products and similar pricing. But that’s where the similarities end.

They sell completely different solutions to different types of customers. One seems like it would have a much faster adoption rate than the other.

After a little investigating we have a hypothesis: both companies rely primarily on the final day of the trial to drive conversions.

In other words, both companies may be missing opportunities to accelerate sales by:

  • creating additional incentives to convert sooner, and
  • continuing to drive conversion after the trial ends.

That’s why we call the “S” curve the $ curve – we see opportunities for using predictive analytics to grow revenue by developing unique conversion incentives for different customers.

All of your customers are exactly the same – so use the same conversion incentive for everyone

See how silly that sounds? Obviously we suggest you do the very opposite.

The trial period isn’t magic – whether it is 14 days, 30 days, 177.6 days, or π days. Your customers will convert at different rates based on who they are and what they do.

Which … drum roll please … takes us to our conclusions …

My advice for accelerating your SaaS sales

Running A/B tests looking for the 1-size-fits all trial period might be a waste of time – there probably isn’t a magic number for all of your customers. Instead, try to optimize your conversion rates based on qualification and behavior.

Pursue post-trial sales

If you simply toss all post-trial customers into your “nurture” campaign you are almost definitely missing opportunities. Predict those most likely to buy based on qualification and behavior – target and pursue them aggressively for at least 90 days after your trial ends.

Or ask us to do it for you.

Create during-trial conversion incentives

The lesson from Observation 1 above is that your customers will convert at different rates. Identify behavior predictors and create incentives to convert them as fast as possible. This is especially true if your conversion curve fits the “S” pattern above.

Yes, we can also do this for you.

Optimize conversions based on value creation – not time

The end of a free trial only serves one purpose – creating purchasing urgency. The best time to create this urgency is soon after the customer generates value from this service. This isn’t as hard as it sounds – and I can prove it to you: try MadKudu for free – we won’t ask you to pay until we start creating real value for you.

Wasn’t that simple?

Epilogue: YOUR business is special

You’re probably wondering … what’s up with Company A?

Screen Shot 2016-02-12 at 5.57.44 PM

Why does it appear to do a lousy job at sales acceleration? Did Company A hire a bunch of monkeys to answer support emails?

monkey (Mixed-Breed between Chimpanzee and Bonobo) playing with a laptop (20 years old) in front of a white background

Nope. No monkeys.

Company A has an awesome product and knows their customers well. Their product is complex and has a slower adoption rate because a customer needs to integrate it into their infrastructure over time – that’s why they have a slow-and-steady adoption rate powered by converting and up-selling freemium users.

What works for them probably won’t work for you.

Every SaaS business is different and special, especially yours. Predictive analytics isn’t magic or a robotic solution – just the math we use to amplify what is already special about you.

Want us to write more posts like this?

We love doing posts like this but it also takes us a HUGE amount of time to look through and interpret the data. If you want us to write more like these please share it on Twitter and vote for it on Growth Hackers.

photo credit: Thalo Porter

 

Use predictive analytics to reduce churn by 20% in 2 days – with 3rd-grade math

Most SaaS companies have 3 misconceptions about churn:

  1. They don’t realize how much churn is costing them.
  2. They think they know why customers churn.
  3. They think predicting churn with data is too hard.

If you’re not using predictive analytics to prevent churn this hack will help reduce your churn by about 20%. It takes about 2 days of work over a few weeks and you can do it in Microsoft Excel.

We used similar techniques to help Codeship retain 72% of their at-risk users.

 

Download the spreadsheet to follow the example below.

You need to predict churn with data

Your customers cancel for lots of different reasons. Projects get scrapped. Users get stuck and bail. The key user takes a sabbatical to breed champion goldfish.

Quite often you can intervene before this happens and prevent it – but the primary predictors of churn are not always obvious.

For instance many SaaS marketers assume last_login_at > 30 days ago predicts churn. We almost always identify better predictors such as changing patterns in user behavior.

Let me re-phrase this point a little stronger:

If you’re not looking at data to predict churn you are almost definitely missing the fastest, easiest way to increase your MRR.

Why this hack is effective

You don’t need a data scientist. Or developer time.

As long as you have access to metrics in Mixpanel, Intercom, etc. even junior members of your marketing team can do it.

Credit card companies invest massively in predicting churn because slight improvements generate millions of dollars. You’re not Capital One – you’re a SaaS company. You don’t need know what “entropy” is to start predicting churn.

You don’t need need statistics

Can you add? This the only math skill you need. There is one equation but we’ve already put it into the spreadsheet for you.

If addition is too complex consider outsourcing to a 3rd-grader. They’ll work for peanuts (or at least cookies).

The results are immediately actionable

We’re going to start with the data you already have in your analytics or marketing automation platform – so you can use the results to send churn-prevention emails or generate alerts for your sales team.

Step-by-Step: find the best predictors of customer churn

Download the spreadsheet

Click here to download.

The examples are easier to understand if you spend a few minutes looking at the spreadsheet. I break down each step below.

PR Power! – our example company

I’m going to walk you through each step using examples from a fictitious SaaS startup called PR Power! we introduced in a previous post.

PR Power! helps media managers in mid-sized businesses do better PR by generating targeted media lists. Customers pay $50-$5,000/month after a free trial. Marketing Mark, the CMO, is charged with reducing monthly churn from 5% to 4%.

Step 1 – Identify predictors of churn

Try to identify predictable reasons why customers cancel.

Mark’s team spent a few hours looking at the last 20 customers who canceled and identified a few predictors. He also interviewed the sales and customer success teams about these customers.

They came up with the following events that are likely to predict why a customer cancels an account with PR Power!

Champion departs – Usually PR manager leaves the customer’s company.

Project canceled – Customer signed up for a specific PR campaign and then decides not to run the campaign.

No journalists – Customer can’t find a good journalist in PR Power! to cover a story.

Support fails – Customer contacts support a few times and the problem isn’t solved – usually indicated by support tickets open a long time.

Stale list – Customer’s media list is less useful because journalists no longer available or active.

Step 2 – Translate the churn predictors to data rules – or eliminate them

Mark’s team took these qualitative events and tried to identify existing data in Mixpanel that might predict them. 3 were straightforward 2 took a bit of investigating.

No journalists required identifying customers who had searched for journalists but didn’t add them to the media list.

Support fails was simply too hard – the support desk data on tickets isn’t in Mixpanel so they decided to skip it.

Step 3 – Count the occurrences of each predictor

Mark put the predictors at the top of his spreadsheet and identified every customer who matched a data rule yesterday.

For instance, User 80374 last_login_at > 30 days ago is TRUE so he entered a 1 for Project canceled.

Step 4 – Track every customer who churns until you hit 100

Mark adds a “Canceled?” column to the spreadsheet. Each day he identifies every customer who cancels until 100 customers cancel. This takes 2 ½ weeks.

Step 5 – Count the matching events for each predictor

Now for the 3rd-grad math …

For each predictor, count every customer where the churn predictor is TRUE and the customer canceled.

matches

Mark starts with the Project canceled rule and counts the following

Number of times last_login > 30 days ago is TRUE and YES, the customer canceled.

For instance, customer 80374 and 89766 fit this criteria. He counts 22 instances.

Step 6 – Enter the results into the spreadsheet

Enter the total in the appropriate block of the 3×3 matrix to calculate the Prediction Score (This is implementation of the Phi coefficient).

Mark enters 22 and calculates Prediction Score for Project canceled at 0.009

Step 7 – Identify the biggest predictors of churn

Rules with the higher Prediction Score are better predictors of churn.

Mark compares the Prediction Score for each rule and sees an obvious pattern.

results

Two observations immediately jump out at Mark:

First, last_login_at > 30 days ago doesn’t tell him much about Project canceled. Since PR Power! has long-term customers who use the product periodically this isn’t surprising.

Second, No journalists is the clear winner. In hindsight, this makes sense – customers who try to find a journalist and can’t are getting no value from the product.

Step 8 – Take steps to prevent churn

Mark creates 2 rules in Mixpanel for the No journalists predictor.

Small accounts

When a customer has total_searches > 5 within last 30 days AND media_list_updated_at > 30 days ago Mark creates an auto-message inviting a customer to watch a webinar on “How to search for a journalist”.

Large Accounts

When a customer has total_searches > 5 within last 30 days AND media_list_updated_at > 30 days ago Mark creates an alert for the sales team to notify them about a customer at risk for churning.

An easier way – ask us to do this for you

You don’t need even need 3rd grade math.

Just take a free trial of MadKudu and let us run these calculations for you.

Cancel anytime if you don’t like it – keep whatever you learn and all the money you make from reducing your churn.

 

Want to learn more? Sign up for our new course.

 

Photo credit: Rodger Evans

How I teach SaaS marketers to accelerate deals

Forbes just released a study confirming what we’re hearing from SaaS CMOs:

78% [of B2B Marketers] see B2B marketings’ role expanding from demand generation to deal acceleration.

In SaaS companies “deal acceleration” means arming the inside sales teams with better information about customers:

  • Improving Marketing Qualified Lead (MQL) quality
  • Predicting when customers are about to churn
  • Providing sales with real-time information about what customers are doing in the product

I’m covering all topics in our new course. In this post I’ll tackle MQLs.

Is your SaaS marketing team ready for this shift?

Do you measure the quality of Marketing Qualified Leads (MQLs)?

Don’t worry, you’re not alone.

Most SaaS CMOs don’t measure and track the effectiveness of their MQLs. In this post we’ll show you how to use a single metric – the MQL Performance Score – to track MQL quality and grow your SaaS revenue.

Why you should care about MQL “quality”

When we interview our SaaS customers about their marketing and sales workflow we usually find sophisticated marketing automation systems and very basic MQL generation systems.

For instance, a SaaS marketing team may “just tag every lead in Salesforce as ‘marketing qualified’ if the trial customer finishes signing up”. We usually discover the following problems:

CMOs have no visibility into how sales uses MQLs

The CMOs don’t know if sales treats MQL differently or even uses them at all. Some sales reps don’t even know what “marketing qualified” means – much less what to do about it.

Sales believes marketing leads “don’t convert”

Sales may use MQLs in ways marketing never expected.

For instance, a rep may tag every MQL as a “Sales Accepted Lead” under an incorrect assumption that someone in marketing already reviewed them. The rep engages many leads who never buy and concludes MQLs “don’t convert”.

CMOs have no feedback loop for improving sales support

Should marketing send sales more MQLs? Fewer? Should marketing supplement Salesforce with key actions the customer took in the product? Did our latest update to the MQL scoring rules improve or reduce MQL quality?

We suggest using a single metric – the MQL Performance Score – to track MQL quality.

Your MQL Performance Score

Every day you run a set of business rules that identifies “Marketing Qualified” leads in your CRM (e.g. Nutshell, Salesforce, or Pipedrive…). Your sales team identifies those most likely to buy and close them.

Your CRM also contains many other leads – what we call “non-MQL” leads – from trial customers, third-party sources, webinars, “contact” forms, etc.

A percentage MQLs convert to paying customers and percentage non-MQLs convert to paying customers.

In high-volume SaaS companies we expect (hope?) that MQLs convert at a higher percentage – if not, something is probably wrong.

The easiest way to measure MQL performance is to calculate your MQL Performance Score:

MQL Performance Score

Here’s how you do it.

Step-by-Step: How to calculate your MQL Performance Score

If you can use Excel and know 5th-grade math you have all of the tools you need. The practical challenge is getting and cleaning up the data – especially since the data is in your CRM and not the marketing stack.

Download a copy of the spreadsheet used in this post.

Step 1 – Break your leads into cohorts

Breaking your data into cohorts helps identify trends and reduces the impact of data anomalies. We suggest starting with monthly cohorts – that is, collect all leads who signed up in a given month and track their progress through the sales funnel over the next several months.

For each month gather the total number of MQL and non-MQL leads. Set up your spreadsheet as follows:

step1

In October 17,000 new leads were added to Salesforce. We broke them into 2,000 MQL leads and 15,000 non-MQL leads which we entered into Column C.

Step 2 – Count the leads in each sales workflow step

Create a column for each step in your sales workflow and plug in the number of leads.

Step 2

(click the image above to see a bigger one or download a copy)

Since your workflow is probably different I’ll walk through each column during October 2015 for the MQLs.

In October of 2015 2,000 MQLs were added to the CRM. Sales accepted (SALs) 440 of these leads (Column E). Sales contacted 396 (Column H) of these leads and 71 of them responded (Column K). Sales qualified (SQL) 66 (Column N) as likely buyers and 46 (Column Q) bought the product.

Step 3 – Calculate the percentage that converts in each step

Calculate the conversion rates for each column you created in Step 2.

Step 3

In October 22% (Column F) of MQLs were accepted by Sales. We calculated by dividing SAL count (Column E) by new MQLs (Column C).

Calculate this conversion percentage for Columns I, L, and O.

Step 4 – Calculate the MQL and non-MQL conversion percentage

Calculate the percentage of MQLs and non-MQLs that convert into paying customers.

Step 4

(Columns E-P are hidden)

In October 2.3% (Column R) of MQLs converted to paying customers (Column Q/Column C).

Step 5 – Calculate the MQL Performance Score for each cohort

Now calculate how much better MQLs performed relative to non-MQLs for each cohort.

Step 5

In October an MQL was 3.8 (Column T) times likely to convert than an non-MQL (2.3% / .6%)

How to use your MQL Performance Score

Getting insight into how sales uses MQLs

Looking at our complete spreadsheet above already raises some questions.

Analysis

What happened in December? Did the sales and marketing team drink too much egg nog at the Holiday party? Marketing only generated 400 MQLs and sales only accepted 300 non-MQLs. This looks suspiciously like a data problem.

Did November provide an example of how we can grow faster? It looks like the sales team paid more attention to MQLs in November. A higher percentage were accepted, contacted, and converted. Did we run a unique campaign? Did a particular sales rep choose to focus on MQLs? Further investigation is needed.

Measuring the impact of changes

Tracking MQL Performance Score allows you to systematically test and measure changes to your campaigns, products, and scoring rules.

Benchmarking your SaaS marketing team against competitors

Unfortunately we don’t yet have enough data to give you a good benchmark – obviously there are tons of variables. An expensive, enterprise SaaS product will have a lower MQL Performance Score than one that sells for $10/month.

For our high-volume SaaS customers we are seeing MQL Performance Scores of 3-6.

And … last but definitely not least … evaluating how much more $$$$$ MadKudu is making for you

Seriously – just sign up for a free trial of MadKudu – we’ll calculate your MQL Performance Score and show you how to improve it.

You have absolutely nothing to lose. You won’t have to pay us a dime until we prove how much more we can grow your SaaS revenue.

 

Want to learn more? Sign up for our new course.

How SaaS CMOs use customer personas to generate better sales leads

Here’s a quiz.

The top challenge facing SaaS CMOs is …

A. Improving marketing automation.
B. Finding more leads.
C. Generating more consistent Marketing Qualified Leads (MQLs) for inside sales.

If you read marketing blogs you probably think ‘A‘ is correct. But if you work with SaaS marketing teams you’ll quickly discover that for most of them ‘C‘ is the biggest challenge. The real work starts rather than it ends once you’ve generated leads.

Why? Just follow the money. Most SaaS companies are trying to close bigger accounts. Marketing automation is great for incrementally improving revenue, but winning major accounts still takes sales. Under pressure to close bigger deals, the sales teams are demanding more consistent, higher-quality MQLs from the CMO.

So how can CMOs generate better MQLs? Well, we answer that exhaustively in our new course.

In this post we cover one part of the answer – how to improve the feedback and communication between sales and marketing teams using Customer Personas. We’ll walk you through a step-by-step example using a fictitious company called PR Power!

Meet PR Power!

PR Power! helps media managers in mid-sized businesses do better PR by generating targeted media lists. Customers pay $50-$5,000/month after a free trial.

CMO Marketing Mark has been building the company’s marketing funnel and automation for 2 years. VP of Sales Selling Sandra just started building the inside sales team and asked Marketing Mark to post qualified leads into Salesforce.

Marketing begins qualifying leads for sales

Marketing Mark and Selling Sandra came up with a workflow which can be simplified as:

Marketing Mark agreed to identify the most promising trial customers (MQLs) and to pass them along to Selling Sandra’s inside sales team. Sales agreed to review the leads and accept (SALs) those most likely to buy.

Marketing Mark’s team spent months developing the business logic to support this process. They added simple scoring rules such as “disqualify any students who sign up with a .edu email address”. After a lot of late nights they got the MQL generation process going.

It was a wonderful plan … until …

For the first few months everything worked as planned. Selling Sandra’s team started engaging the leads and paid conversions grew by 30%. Yipeeeee!

Then reality hit…

…the CEO decided to focus the company on bigger enterprise deals…
……the product changed to support larger customers…
………Marketing Mark’s team struggled to keep scoring rules updated…
…………and Selling Sandra (under pressure) started generating her own leads.

One day Marketing Mark realizes he’s investing a ton of resources generating MQLs that sales teams don’t even use. He doesn’t know why.

The Meeting: “Why isn’t sales using the leads from marketing??”

Marketing Mark calls a meeting with sales to discuss.

Marketing Mark: “I know inside sales is under a lot of pressure to grow revenue. We want to do our part. Last month we sent you 300 MQLs and you only accepted 3 as SALs. Why?”

Selling Sandra: “Wow, 3? That’s 3 more than I expected. Your leads suck and I don’t want my team to waste time calling them.”

Marketing Mark: “Ok … I need a little bit more feedback than ‘sucks’. Believe or not we don’t have a ‘suckiness’ customer attribute in our database.”

Selling Sandra: “Last month I called one of the higher scored MQLs you sent me. I spent 3 hours playing phone tag with some guy who turned out to be a student doing a class project. That’s what I mean by ‘sucks’.”

Marketing Mark thanks everyone for their time and promises to explore the issue further.

If only we had a “suckiness” customer attribute…

Marketing retraced Selling Sandra’s ‘student’ lead and discovered the trial customer wasn’t using a .edu email address – so the lead wasn’t scored as a student and became an MQL.

Unfortunately the marketing teams feels like they’re being blamed. As a joke someone writes on the whiteboard:

sucky

Since there is no “suckiness” attribute the team has to figure out ways to identify the attributes and behaviors of poor leads.

This is really hard without good collaboration & feedback from sales. This situation is so common because marketing and sales people think about customers differently.

Marketing thinks data. Sales thinks people.

SaaS marketers think in terms of events, attributes, cohorts. Sales teams think in terms of people. They think about customers differently and use different language.

Unfortunately this difference can cause the problems like those at PR Power!: Sales has no tools for providing feedback on MQL quality in a way that marketing can translate to data and business logic.

Customer personas are one tool for solving this problem.

Customer personas create a common language for sales and marketing

Personas describe a customer in a way that can be mapped to attributes and behaviors. They are simple to make, easy to understand, and easily changed.

Here is Customer Persona template we created for use in SaaS sales funnels:

sales_persona

Step-by-Step Example of using personas to generate better mqls

1. identify leads Sales doesn’t accept and group them into common archetypes

Why is Selling Sandra so unhappy with the MQLs? Because top-performing sales people are busy and want fewer, high-quality leads. The fastest way to improve MQL consistency is to eliminate poor leads.

Marketing Mark interviewed sales reps and learned that sales didn’t want to waste time talking to students, startups or freelancers since these customers – although very active – were unlikely to become larger accounts.

2. Generate personas for each group of leads.

Mark created customer personas for students, startups, and freelancers based on customer attributes and behaviors. Using our template above, here is the “Student Sammy” customer persona:

sammy

Tips:

  • Keep them simple.
  • Silly, descriptive names are easier to remember. e.g. “Student Sammy”, “Startup Steve”, “Freelancer Freddie”.
  • Don’t go for perfect.

3. Update MQL generation rules based on the personas

Instead of using a simplistic business rule like “disqualify leads with .edu addresses”, Mark refines his business logic based on all of Student Sammy’s behaviors and attributes.

4. Ask sales reps for immediate feedback using the personas

Marketing Mark asks all of the reps to”let me know if we send you any Student Sammy’s”. Now Mark is in a position to get contextual feedback.

Tips:

  • Inside sales reps often have to be prompted for feedback.
  • Post the customer personas on a wall where sales reps can see them. Funny pictures help.

5. Refine and update the personas over time

Fast-growing SaaS teams update their personas every 6-10 weeks. Often they are too general and need to be sub-divided.

When sales reps wanted to contact graduate students Mark split the Student Sammy into “Undergrad Ulf” and “MBA Mickey”.

Want to learn more? Let’s talk!

Personals are just the 1st step in creating high-quality MQLs. The CMO’s marketing team will need to analyze data and develop predictive models for attributes and behaviors that map onto the customer personas.

Lucky for you … we’re here to help. Sign up here and you’ll be on your way for turning the VP of sales into your best friend.

 

Photo credit:Gabriel Cabral

Who owns SaaS trial conversions?

 

Letting customers try your product before buying is becoming a standard practice.

Free trials are now more and more common. For example, we analyzed a sample of 41 Techstars SaaS companies and found that 77% of those companies offered a free trial.

Well known B2B SaaS companies like Salesforce, Zendesk, LinkedIn, and HubSpot work with this model and are defining customer expectations in the B2B world.

Free trials are popular for a reason. They are a great sales tool. They allow you to “soft sell”. They make the ask smaller. They reduce the perceived risk in the purchase decision. They are similar to the free return policy now offered by almost every retail store.

Many companies do quite poorly at optimizing free trial conversions.

We’ve always been surprised by the amount of effort put into adding more leads to the top of the funnel in comparison to how much is done to convert those leads.

Many of the companies we work with have trial conversion rates ranging from 1 to 15%. In other words, 85 to 99% of acquired leads go down the sink. Even if you assumed that 70% of those signups are not potential customers, it still leaves lots of room for improvement.

Increasing a trial conversion rate from 3% to 4% means reducing customer acquisition cost by 25%!

The free trial stage is the most complex of the customer journey.

Most stages of the customer journey have one clear owner. For example, marketing is responsible for bringing traffic to the website and converting this traffic into leads.

The free trial stage is more complex. It involves almost every department:

  • Marketing: set up email drip campaigns to guide and convert trial users, experiment with discounts and pricing.
  • Customer Success: onboard customers and coach larger accounts to become successful.
  • Sales: explain the value prop, give demos, help customers pick the right plan, negotiate contracts.
  • Product: identify friction in the product, improve product user experience, add missing features.

The lack of a dedicated owner results in sub-optimal trial conversion rate

Are you familiar with the business fable “the chicken and the pig“?

The trial stage often has lots of chickens but it rarely has an assigned pig.

Everyone has a critical role to play there but the contributions are usually tactical and of limited impact.

The best companies I have worked with assign a strong owner dedicated to optimizing this stage the same way, let’s say, a website is optimized: gather data, make hypotheses, test, learn, implement, iterate.

“Okay, okay… who should own trial conversations then?”

I have seen different configurations. Most of them depend on the type of SaaS businesses: high-volume versus  high-touch.

In high volume SaaS companies, the CMO usually owns trial conversions. Or more accurately, marketing owns the conversion of self-service leads (usually defined as signups from companies likely to buy a small plan) while sales owns the conversion of enterprise leads (signups from large companies). Marketing works closely with product to test different discounts, pricing, and plans. And they work with customer success to implement effective email drip campaigns.

In high-touch SaaS companies, the sales team assisted by the customer success team tends to own trial conversions.

Having a CRO (Chief Revenue Officer) is a new and growing trend, particularly in the Silicon Valley. The CRO oversees and “optimizes the entire customer experience with the aim of increasing revenue”.

Let me ask you the question now. In your SaaS organization, who owns trial conversions? How is it working? Share your feedback and thoughts on twitter or email!