I have a very boring approach to marketing.
Unlike the vast majority of people I meet in the space, who believe that 95% of marketing is creativity and ideology, and then you just do it and you move on to the next genius idea. I'm solely focused on the execution.
The idea is 10%, but the other 90% percent is understanding that every marketing action, every exercise, every activity that you do essentially has an input, an activity and an output.
This ranges from PPC, where you have input which is audience plus dollars, activities - the ads that you run on the landing pages you build and your output is X, Y and Z leads. And that's the easy understanding.
But it's the same when you're doing PR, and I've had amazing successes in some companies with creating a hell of a lot of noise. But still the equation’s the same. On one side is the audience, the journalists and then there’s the PR activity that you're doing, and on the other is the output, which is getting hundreds of people to write about you.
Now, when you follow that process, as boring as it sounds, you sometimes realize people have been doing it wrong for so long because they don't understand what the input is, what the activity is, and what the output is.
My previous company Feedvisor, for those who don't know, is an Amazon repricer. Which essentially means we sold to people who sell on Amazon. The West’s largest online marketplace, where hundreds of billions of dollars worth of sales are made -- and primarily made by sellers. I.e. Individuals that sell on Amazon.
Amazon themselves sell 5-10% of the items, and what they want to do is expand and offer products to everyone else, but they let other people sell through Amazon.
And when you're selling on Amazon, one of the biggest questions that everyone always has is, how do I set my prices? If I'm too high I'm not going to make any sales, if I'm too low I'm not going to make any profit, so any repricer essentially balances your profit vs. sales to maximize your end results.
Back then we were the very first people, we coined the phrase "Algorithmic Repricing", where everyone else were “Rule Based”, and this leads me to something else that I always try to do, which is define a space and own it.
Don't fall into other people's spaces, because then you're fighting them.
Define the category, and then your content and your communication strategy should be educating people about that category.
You happen to be the leader in that category, maybe you're the only one in that category because you made it, but then you can push your agenda without pushing your company.
One of the things that I noticed very early on is there was something called "The Amazon buy books", which is that small box that says "Buy now", and you click it, and every time you click it it makes a sale and it gives it to one of the sellers on Amazon.
Now, how it determines which seller on Amazon it goes to, was and still is, a highly kept secret. Amazon doesn't want you to know how it determines who wins this because it's like Google, you just hit the right points and all of a sudden you'll make all the money.
They want to avoid it the same way Google SEO algorithm is one of the best kept secrets in the world. So one of the things that we noticed is that a lot of people are asking questions on the Buy-books, but not a lot of people knew a lot about it.
So we went down the route of the best way to make yourself into an expert in anything, is to tell people that you're an expert on that thing.
Then we started researching and learning about the Buy-books, and very soon, in a matter of weeks we read every comment and every article, and we realized we knew more than anyone else in this field. So we made a big fuss about being the Buy-books experts. We were announcing that we were going to launch "The bible", we were going to bust it open.
Beforehand we did something very clever, we reached out to journalists, we reached out to opinion leaders, bloggers, big medium and small, everyone who'd ever written about the Buy-books and basically said:
"Hey, we saw your post three months ago about the Buy-books, it was inspiring, in fact it was so good that it gave us an excuse to spend months researching that idea you had and you wrote a comment about it somewhere, and now we've actually written the manuscript of the Buy-books bible. Two things, we would like to credit you in the back because you've given us so much, and second of all we'd love it if you could read through this".
And we sent this out to about 120-130 different experts in the field. I think that as a small startup that no one had heard of, we had three or four people replying. Those who replied were nice, they gave us genuinely interesting comments, but the truth is that by that point we'd already finished it, copy-written it and we were way into the design stage.
And then a few weeks later, when we launched it, as a PDF, to every single one of those experts that we first asked for help with.
First of all we said, "Thank you so much, you've been overwhelmingly generous and you've helped us so much, and we've really put together something which we could never have done without your specific genius, we've put your name at the back as a contributor, so you're welcome to tell people that you contributed to the Buy-books bible, and we'd love it if you could share a link to our landing page which is here".
And we had something like 80-85 of these experts in the field, all tweeted, all shouted, all at the same day, we gave it to them all a couple of days early with an exclusive and said, "We're not sharing this with anyone else, but you know, you wrote the thing, you contributed so much, we're launching it next week". And we had something like 6,000 downloads in the first 48 hours.
My attitude is, you know there's this expression, throw things to the wall and if it sticks, then you know it's good? I have two caveats to that.
Number one, it has to reach the wall.
If you're going to spend 50$ on Google AdWords and then tell me that Google AdWords doesn't work for your category or industry, it hasn't reached the wall.
Second thing is, and this is sometimes for senior staff and for CEO's maybe a little bit harder, is once it’s hit the wall and it sticks, go for it.
When you are in a stage where you're doing trial and error, and you want to see what works and you're testing a 100 things, the most important thing is, fail well and fail quickly.
If it works, then you have to jump in and you have to really commit, and often you'll find that when things aren't looking that good, when budgets are a little bit tight, people do too many things in a way, because it's excusing them.
Then they can say, "Well we tried a whole lot of things and it didn't work", but if something did work then, own it, go for it and double down.
So normally, the best way I ideate new content in a new space is I start reading, researching and find the things that people don't see enough about but are interested in.
I'll give you an example - at LawGeex we worked on something called "The Legal Tech Buyers’ Guide", which was a very similar paper to what we did at Feedvisor.
It was 150 pages long and it listed every single legal technology. By the way, all our competitors were there, with their phone number and a link to their website. And the reason we did this is because there were so many people online saying, "I want to start using legal technology, I don't know what to buy". So when you say, after spending a week, two weeks, a month, kind of living in your industry, speaking to your customers and seeing what the experts are saying and what they're not. You soon get a feeling of where the niches that you can ask your questions.
And by the way, with the Legal Tech Buyers’ Guide we did something similar, and we did it without competitors. We said to all our competitors, all our competition, everyone, every legal technology in this space, "We're going to list you in the new guide, feel free to send us the description of yourselves, we'll add it and then we'll share that out".
Obviously, when we shared that out, all of our competitors shared our guide, because we listed them all in the top three of this space, or top five of this space, which was incredibly useful. And it gave us the opportunity as the guys that wrote the guide, to define our competitors exactly as we wanted to define them.
So, if we were let's say, looking at enterprises, and we had a very similar competitor, we would write, "This is a great piece of software, great for checking contracts for smaller companies, recently they've been moving into enterprise companies and we all wish them the best of luck".
This allowed us to own the conversation and position our competition exactly where we wanted them to be.
And where you're positioning the market, I mean, one of the best things that came out of those, that Gartner themselves created a category that we define for ourselves and we wrote in the buyers’ guide.
We told them where we wanted to be seen, and because nobody else was defining the categories and was defining the structure, Gartner said "Well, if LawGeex have done this already, we'll take their definitions and we'll put them where they say they are". Which was very useful.
Creating a category is also something we did at Clicktale.
We made up the term "Customer Experience Analytics", when everyone else was just “Web Analytics.”
We moved towards Customer Experience Analytics, which was then taken over by Adobe and a lot of other people who are... You know, it's now its own real field.
Defining a category is something which, when done well is, one of the best and first tools, pieces of ammunition that you have in your toolbox.
Because when you define a category, three things happen simultaneously.
The industry leaders, the guys that know everything, want to know who you are, because there's a term, there's a buzzword, there's a keyword they haven't heard of before. Therefore they don't want to feel tripped up.
They don't want somebody else to go to them and ask, "Hey have you heard of Customer Experience Analytics", and they're like, "I'm not sure". So when you offer to present to them, you get to do it.
The second thing you get to do when you define a category, and this is very very valuable, is then you can take the role of the evangelist much easier. And you can spend a lot of time telling people why Customer Experience Analytics is important. You don't have to tell people why they should use Clicktale, it's secondary.
You paint the picture of what customer experience should look like, how you get the analytics to match that, and what theoretical software you would need to match the two together.
It allows you to evangelize without sounding like you have an ego.
The third thing, and this is also very very valuable, is once you define the category, you get to define the market.
So you can say who is and who isn't, and it's something which we're working on now. And by saying who you are, and I learnt this from the people I think did it best which is Mailchimp.
They say who they are, by saying what they're not.
So if for example Mailchimp were fighting IBM, they would say, "We're powerful, but not complicated". And therefore without being negative, without throwing shade at anyone, they're saying, other people in this space are complicated...
...and obviously that means the IBM's and SAP's and the scary guys.
You know, they say, "We're powerful, we're not complicated, we're friendly, we're not to be underestimated". I'm not entirely sure how they worded it, but they get to hint as to how you should see everyone else. Which is again very useful, so we did this at Feedvisor again by saying that:
"We are algorithmic, we're not rule-based", and then every repricer by definition was rule-based. And people went, "Wait, one second, what's wrong with rule-based?".
And then we go, "Glad you asked, let me show you". And you get to define the them-and-us. And you can do it without being aggressive, or without making it sound like a battle.
You just say this is them, this is us, these are the constraints that we're working with and therefore choose your path, you can go with the new technology or with the old and more complicated one.
This back and forth positioning came together for me in a really nice way during my time at Lawgeex. One campaign in particular.
AI vs. lawyer was, I think I can say, one of the highlights of my career, both in how much fun I had doing it, and the outcomes at the end of the day, which blew myself and my whole team and the industry away.
Essentially what we did in AI Lawyer, the theory was, we were sitting in a board meeting. I think this was in Q3, and each team was presenting to the board, and the R&D guy stood up and said,
"Actually right now our algorithms, as of this quarter, are more accurate than humans".
And my marketing was shot up, and we said, "Wow wow wow, really?"
And they said yes. We asked if we can test this, can we validate this?
And they were like, "Sure, why not". So from the RND side they weren't seeing this as a leap, but from my marketing point of view this was the turning point that could be essentially another... you know, Google AlphaGo moment.
That's what I was thinking.
So, we spent the next six months, and a lot of credit here has to go to a lot of the teams in LawGeex that really invested. We invested quite a bit in doing this properly, Johnathan Marciano who headed up our communications, he had a baby half way through the process, and I wasn't sure which one of the babies was his true baby. He put in so much effort, and a lot of credit goes to him and the rest of the team there. But we found online independent adjudicators, people that would actually help us put the test together.
We found 20 fully qualified professional lawyers, with experience reviewing contracts, who were willing to go head to head with a machine.
Sure, we did this because we knew we were going to wing.
But, it's more than that. And this is my shining message throughout the whole process, and I'll get back to the process in a second. It's all about giving people the WIIFM, what's in it for them.
So for example to the lawyers we said, "We're going to make you a star. You're going to be the people that fought the machine, if you win you beat the machine...
...if you lose you are at the precipice, you are at the pinnacle of the legal AI revolution, and when this goes global, your name's all over it". And, I'm not saying lawyers have egos, but there's definitely those lawyers that like some publicity.
And the same thing with the adjudicators, we said to them, "Guys we're going to give you THE NEXT THING, we're putting it in your lap". We went to four different universities, we went to Stanford university and three others I can't remember their names, but four big major universities.
We initially went to dozens, but these are the four that said yes, I think Yale may have been one as well, and we basically said to them,
"We want to do this study, we want to do it properly, we don't want marketing fluff, we want to do a proper study, we're essentially giving it to you, we found the lawyers, we've written the AI, we've found the adjudicators, we wrote the test, all I want you to do is check the tests, make sure it's clean, when the results come in we'll send them to you, you can tell us that it's Kosher. And then you can put your stamp on the paper".
And every university wants to put papers out, that is kind of what their WIIFM is, and they want to be on the papers that get in the news because that's how they get grants.
So, we really did do the study properly. We got 20 lawyers, we got our AI here in Israel, we found five NDA's which were made public after the M1 scandal, and we gave those five NDA's to the 20 lawyers, we gave those five NDA's to the AI. And they weren't contracts which we used anytime to train the AI, so it was a true test.
Although obviously we went with a test that we were very very comfortable we'll win. We weren't asking them to negotiate, we weren't asking them to correct, it was just review the contract, highlight discrepancies, find these issues. That was it.
So it was the most basic work, so we're not saying high level stuff. And we put all those outs together, we found out that the lawyers scored an average of 85% accuracy which is good, really good, and we reminded them that in certain states to pass The Bar you need 60%, so nobody can doubt that question.
And then the AI on the same contracts got 94% accuracy. And obviously it took the lawyers an hour or two to do the test, and it took our software all five contracts 13 seconds.
And we took this story huge, and in no point did we ever go to the journalists, and say to them, "End of lawyers", you know, "The profession is coming to an end".
But we fed them almost everything until that line, then we waited for them to ask us if this was the end of the lawyers.
Then we went, "No!", this is the beginning, this is like autopilots for planes, there's no less pilots now then there were a 100 years ago. In fact, there's more pilots.
But we sped the story up, and we spun it as if this is the next big AI landmark, that we haven't been able to break.
Kind of like IBM’s Deep Blue.
You had Kasparov and Deep Blue back in the seventies and eighties. Then you had Google AlphaGo, and now you have AI vs lawyer.
And we fed it to them like that, and we were able to push this so hard, I think quite honestly we didn't sleep for five days until this came out. We pushed this to everyone.
And we got 600 pieces of international major media coverage. I'm talking, obviously the VentureBeat, and the TechCrunch, all covered in Mashable. It was Mashable's main piece because initially we gave them the longest interview, they were on the dot on the day with the full article.
That was picked up by Daily Mail in England, by Popular Mechanics believe it or not. We had everyone from the New York Times to the Washington Post. Last week HBO did a documentary on LawGeex and on the study that we did last year and put it on Vice News.
And for me, each one of these things by themselves would've been worth the five months of effort. But when you combine them, I think that probably fed us enough interest that we could've turned off the marketing machine for 18 months after that.
When I left two months ago, we were still getting leads daily, because people read AI vs. Lawyer. And for me, the kind of icing on the cake is,
Forbes did an article... You know, as these articles are often written. The history of Man vs. Machine. And they showed Kasparov, and they showed Google AlphaGo, and they showed the first robot, and they did the first ones that did the mails, and then like number seven eight or nine was AI vs. Lawyer. "Israeli startup LawGeex, blah blah blah", and it was like, that's it.
Because then you're in the consciousness, then you're in the zeitgeist, a part of the tapestry.
Shmuli Goldberg is a SaaS B2B marketing leader, currently CMO of Identiq, and previously leading marketing at Lawgeex, Feedvisor and Clicktale. With 12 years of experience and four successful startups under his belt, he’s had quite a lot of wisdom to share.
Techie Talkie, the tech marketing podcast is a casual meeting place where the best tech marketers share their most impactful trade secrets and marketing hacks . Its objective is to inspire creativity within the tech marketing space and help marketers rise above the noise. It is hosted by Asaph Shulman, a serial marketer and CMO of Firebolt and our very own Carmel Yoeli.