Siddharth Sharma: Marketing Analytics & Decision Making For A SaaS Startup

Moderator: Siddharth, could we start with your marketing journey in brief?

Siddharth: I started as a techie in Infosys then moved into pre-sales in Infosys only. Post that did my MBA in Marketing because marketing appealed to me the most as it is both art and science. Joined Manipal Global Education where I was working in the presales team of a start-up incubated there. There I also worked in digital marketing where my first project was marketing automation as Zapier wasn’t there at that time and had to write a Cron to pick up data from our CRM and push it to an email system. I fell in love with it. And my love for it continued. Post Manipal, I joined Simplilearn- an ed-tech company where I owned content marketing and marketing automation piece, and grew the marketing automation from $10K/month to ~$100K/month. And during that time everybody in India caught the entrepreneurial bug and wanted to be an entrepreneur. I started my marketing agency and even after 7-8 months, forget breaking even I was not able to get over my rent. Joined in a B2B Marketing role at Clevertap and then Karix and then now Verloop. Been doing marketing all my life and will do marketing mostly all my life. A lot of people think marketing is just one function. I am not demeaning any function. But for other roles, you require a certain set of skills e.g., customer success or accounting. And the only thing I could find equivalent to marketing is a full stack developer because you have to start from budget planning. Towards the end, you have to optimize everything. You have all- content, lead generation, brand, PR and to get all those skills developed in a single person is anyway unicornish. But even to have a department that can execute those skills is very tough. So I find it fascinating how to do that, especially in a startup. How do you overcome barriers? What do you do when you fail? As a marketer you set the right expectations, you show the right data and you own up to your failures. You're not going to be successful every day as a marketer. And marketing is probably the most open and apparent function of everything. You do a campaign and make a mistake and everyone in the company knows about it. You do a typo on our website, somehow your finance person will screenshot it and send it to you at 1 AM.

Moderator: Or you sent the marketing automation email to the wrong list.

Siddharth: (laughs) Yeah. Maybe it goes like - “Hi, first name. How are you?”

Moderator: (laughs) I mean such it's one of those- You have to do it right every day and when it goes wrong once, it is so obvious. Anyways, could you give us a sense of how the marketing is set up at Verloop and draw parallels to other SaaS companies, and even contrast it with B2C?

Siddharth: I’ll start with the difference between B2B and B2C. In B2B, most of the team is structured around content and in B2C it is more focused on performance. Content is mostly separated in B2C. When I started in B2C and I used to believe that demand generation is the linchpin. It finally boiled down to even in B2B, if you have a demand gen engine, it cannot sustain without your content. It will die out or it will just run a very short distance. 

At Verloop, I think of the structure as a production unit and a distribution unit. We have a content team that looks at all forms of content, which is your product page, product content, sales content, marketing content, and every kind of content that comes out of the world from Verloop. Then you have the distribution team which is divided into PR, demand generation, and the SDR team. The PR team focused on getting the word out and uses the help of the content team to take out certain angles or stories that we want to share with the world. The demand team is focused on generating more demand. It is in a nascent state and you can call it lead generation ++. Then we have the brand team which looks at how to leverage all the things that we are producing to go higher on the brand value with our enterprises.

Contrasting it with my previous companies we had different functions set up- content, demand gen, and others but I would say there was a lot of non-structured feedback. But something we have tried to actively try to push in Verloop is structured feedback around the production and distribution units. This is what I've tried to try out here as an org structure.

Moderator: What does demand generation include beyond the actual execution of campaigns? 

Siddharth: Demand generation also includes marketing ops. Because the entire analytics around the content, paid and the other functions have a direct tie back to revenue. And that's the only possible when all my tech systems are working singularly, like the way I want them to work. So demand generation also includes marketing ops for me. 

Moderator: What would brand include?

Siddharth: Brand includes analyst relationships, influencers, and to a certain extent social media as well. Since we are in a very nascent stage, social is part of content. As in when it matures, we will spin it off into a different function. 

Moderator: And for each of these teams, what would be the north star metrics for them? How do you measure that?

 Siddharth: I’ve tried to tie back to Annual Recurring Revenue (ARR) since we are marketing at an early stage company, However, tying back the content production unit back to ARR is quite a stretch. For content, we have broken up into two- 1.) How much you're contributing in terms of how many plugs can you put in your content and tie it back to a revenue model. 2.) How much of your content is having an impact? And this is qualitative. How much of an impact it's having on deal velocity? Is your content being actively shared by inside sales or the SDR team to generate more interest or push the deal faster or at least grow the brand? There is an SEO angle where you track the ranking of a particular piece of content over a time period. So that's how the content team is judged.

Then demand gen is totally about the ARR. PR is about how much share of voice you have been able to generate in a market depending upon the maturity of the tech that market and depending on the maturity of the org in the market. Brand is a share of voice that is completely on your online and how much were you able to generate?

Moderator: How do you measure the share of voice?

Siddharth: Engagement and mentions on my social media feeds or blogs including shares and reactions to my blogs. We measure the share of voice through a tool. Then we also measure how many times it was mentioned in the print as well. Say if a Middle East paper like Khaleej Times carries us. So that does have a huge impact on our viewing. But Khaleej is different from a random local blog. So we look at it in a segregated manner. How many were category-A publications and how many were non-A? So you give category-A a higher rating. So then you average those mentions and calculate the rough share of voice compared to my competitors in the market. This is what we are going behind, and this is how much we want to grow.

Moderator: For the content team, you mentioned the SEO metrics- ranking, page views, and most specifically deal volume. How do you measure if the content is consumed by a user part of a sales process? 

Siddharth: I measure that using HubSpot today. It has a multi-touch attribution system. So I tie it back to my SFDC’s Closed Won stage. And I can pull up reports on whoever has been closed in the past, what are the content pieces or the landing pages that the person consumed. And if you have written a piece 6 months back and it has not featured in any of those, then there is something wrong that you're either not foresighted enough into saying that this kind of content does not work, or we are targeting the wrong set of keywords. And we have to look at what kind of content they are consuming. And this is your primary data. Your secondary data is the user research that happens with our end users. So what kind of content they're consuming? What kind of content would they like? So this is where the product team plays a role for us, where they go to the end-users and talk to them and come back with specific data points.

Moderator: The reports on HubSpot and Salesforce, is that like a dashboard in HubSpot that you've created or the marketing ops team creates them? 

Siddharth: Right now we are a small lean team. Marketing ops is shared between me and one other person. And we have a dashboard for the entire team to view to know what kind of content is consumed, etc.

Moderator: How big is the full marketing team?

Siddharth: Eight people across. 5 in content and one each in demand gen, PR, and marketing ops.

Moderator: That is a lean team. How do you judge the quality of the content?

Siddharth: We refer to data for that. The problem comes down to what is trending as of now as well. E.g., the WhatsApp chatbot is trending today. Six months down the lane, let's say WeChat comes to India and now my WhatsApp blogs that I've created will not make a difference to anyone. So there is a right span of each content piece that we have. And we judge it by understanding whether it has been consumed by 50-60% of deals during its lifespan. But if it's not even consumed once in the last six months then we ask why did we write the blog? So we have to look at what was the intention of writing that blog? Who are we writing the blog for? And what are the keywords you're targeting? Because we obviously cannot ignore SEO.

So we start from that thought process and then we come back and say that the SEO keyword is WhatsApp and customer support automation. Now we need to understand for whom are we writing this blog? The CTOs. What do the CTOs want to know about WhatsApp though? The CTO would not want to know the pricing and the slabs for WhatsApp. That is a business question. They would like to know technicalities? And you got these out on WhatsApp and legalities of WhatsApp. What kind of tech is supported on WhatsApp? What type of APIs are supported by WhatsApp? So then you would want to write about the implications of WhatsApp on your tech stack.

Moderator: Do you also measure things like how much time do they spend on that blog?

Siddharth: Yes, content marketers measure those metrics to judge them at their level. If I have a question at my level, I take it by the number of people who are impacted by this content. If I have a question on any piece of content, then they come back and say- the time spent on a page or a blog was very high due to which our domain authority increased due to which we have an indirect impact on our brand and our direct search. So these are content marketers metrics whereas my metrics are how many people were impacted.

Moderator: Got it. In terms of measurability of marketing and how comfortable are you with the way it is right now at Verloop and drawing parallels to CleverTap given that it is a slightly more mature model? What would be your assessment of that?

Siddharth: At start-ups, a lot of people try to own up only until a certain state of the funnel like MQL or SQLs. However what CleverTap has taught me is that you have to own a piece of revenue, irrespective of the stage you are in. However, that's not the only Northstar metric you can focus on because if you focus on that, then you'll lose out on brand and your essence of being a brand. So we also try to focus on other metrics, which are slightly more brand-oriented, which are deal velocity and share of voice. If my average velocity in the last six months cohort is 28 days, in the next six months my target would be to do 10% faster. Then I would want to do it somewhere around 25 days. How can we do that? We can do some email, podcast, webinar, etc campaigns around them.

We can do email campaigns and we do podcast campaigns. Now are these measurable? Not that much. But do these create an impact? Definitely. So what we do is we try to measure 4 things- 1.) Marketing sourced ARR, 2.) Marketing impacted ARR which is during the sales process was there any marketing touchpoint which impacted the deal, 3.) Share of voice which will tell you how much is your brand value and the 4.) Deal velocity

Moderator: Do you also look at an increase in deal value in addition to the deal velocity?

Siddharth: I think the deal value is a combination of things. In B2B, more value a product provides, the deal value increases. The more value a product can provide a person, the more happy a person would be to pay for it. Say if my product doesn’t have an SSO vs my product has an SSO- an enterprise would more likely throw a lot of money at a product that has an SSO.

For somebody who doesn't have an SSO, no matter what marketing you do on it, that money would never come to your pocket. And so it's sort of like a marketing/product problem. Marketing can say that they won’t qualify or market to companies below the 200 or 300 employees size range. But owning a 20% increase in ARR alone by marketing is a bit of a stretch for a startup. All of this is specific for a startup. 

Moderator: Got it. What role, if any, does analytics play in marketing any experience that you've had where analytics has moved the needle.

Siddharth: I'm a big believer in the fact that what does not get measured gets ignored. It is like losing weight- you don't measure it weekly, you don’t care about losing weight. When you see you are not losing weight, you are motivated to not have that “Chhola Bhatura”. It is a human tendency to optimize or scale a number. If you see an error rate of 90%, tomorrow you would want the error rate to be at 89%. Because you want to show that I have moved the needle. It's a human tendency. And therefore I think analytics is a great way to hack that emotion. Obviously, it leads to much better performances because first people now know what they are being measured on and second what they have to optimize. People are not hay-wired and not just running around to know what they want to aim for. When everybody also looks at data on a daily, weekly, and monthly basis, any pivot or intervention can happen pre-blow up. So analytics helps you predict problems that might come in and makes sure that your team is focused on that track. I’ll quote one example from my previous organization where we were looking at publishing blogs. ‘What is a good number of blogs to be published per week?’ is the kind of question your CEO would ask. Because we can build a team of freelancers and in-house and publish any number of blogs. At an optimal level, you can write a hundred. But the question is should you write a hundred? So how do you do that now? This is where analytics helped. Firstly, what is the end goal that we want from content? E.g., we want a 15% or 20% month on month or quarter on quarter increase in organic traffic. So if we do 10 blogs today, based on the historical data and industry benchmarks, it would take 5-6 months for it to rank. Now, if we have to go to the third page to get these many views on this page, this is how much organic traffic we grow. So how many do we need to write today in this month to grow by this much by the end of six months? Then you backward calculate to get to a number. We need to write, say 15 or 20 blogs depending upon the niche or the area you are in. Next, is that number feasible to do it in the budget that you have? Maybe you have only one person now and every freelancer charges exorbitantly for writing a single article. Now is 20 a number that you want to go for? Maybe we can write 5 or 6 and then can we use secondary channels like paid campaigns or email campaigns to drive more and more visitors to our website. Now you say this is what we’ll go for. Only analytics can help you play that role. Otherwise, it's your word, experience, and thought process versus somebody’s experience, thought process, and mindset. 

Moderator: In this specific example, it's not just pure analytics that a dashboard can give you. But you take historical data from some dashboards based and then apply judgment on top.

Siddharth: Right. You also need business context. I could be an early-stage startup or a seed startup or I might be a startup that is funded super well that I have $500 million. If I have that much money in my bag, I would not mind spending on an army of freelancers to generate 50 blogs a month. But if I am an early stage bootstrapped startup where I have $1-2 million in my kitty, I would want to optimize for which are the best blogs or best pot shots I can take so that I optimize my growth. Because there are a lot of players in the market for every keyword that you can possibly think of. Now, if you want to compete with somebody for a keyword like ‘buy sports shoes’. It is super difficult organically. But say a keyword like ‘orange sports shoes for men in the age of forties’. Now, while this will not get you a lot of traffic, it’ll get you a certain set of traffic say 30-40% if you’ll write a decent piece of content because you’ll rank faster. So you go for moonshots and you also go for a fixed ROI.

Moderator: Any examples from paid marketing as well where you've been able to use data to sign up optimized campaigns?

Siddharth: I’ll take an example of Linkedin. LinkedIn for far longer has a bad dashboard. Their campaign manager requires a lot of re-work. LinkedIn has lead gen forms and when you try to import those forms, a lot of data gets lost. Especially like IP country, the timing of the form, etc because the Zapier API might work after 1-2 hours. So you don't know when things are happening. But when HubSpot came out with their LinkedIn ad integration, I was excited as I could predict with a certain confidence that a campaign actually worked for me or not. And then campaigns that were not going to convert into opportunities or not getting any lead, I was able to shut those campaigns or look at the messaging.

Moderator:  Has multi-touch attribution been a problem either at CleverTap or Verloop and how have you tackled it?

Siddharth: Attribution is always a problem. From an analyst’s perspective, I want to give the true picture to my organization whether a deal was sourced by a salesperson however he did 5 to 6 searches on the web before he even replied to sales. Then I want to know exactly what steps the user took if that can be scaled or something that cannot be skipped and show to my organization what it takes to practically crack that deal. This is a simplistic way of telling. But you won’t be able to track a podcast touch because that doesn’t get tracked into your CRM. For example, you don't mention that you got the Delivery Head of Swiggy to your podcast. It doesn't happen and you miss out on that event. So marketing attribution will remain a challenge because of data. However, I'm seeing new tools come up in the market that are solving that problem from day one. I think it's a fantastic boon for marketers. And I think all CMOs should pay close attention to attribution while planning and budgeting.

Moderator: Got it. At Verloop, do you look at any specific attribution model or is it still too early for Verloop to focus on attribution?

Siddharth: We look at the first and last touch and for multi-touch, we look at the impact. Now, the impact can only be measured by multi-touch and we only use it for content. We don't use it for the ARR. E.g., we don’t track what % of ARR was contributed by marketing. Because like you said it is kind of early for us. However, in a couple of years, we should be able to talk about a pipeline, multiple touchpoints, and how much of the touchpoint belongs to marketing.

Moderator: In one of the LinkedIn posts, you had mentioned you love processes over outcomes and have like core frameworks for decision-making. I've seen a couple of tissue paper drawings that you put up on LinkedIn as well. Firstly, could you highlight 2-3 frameworks that you use for evaluating marketing? And secondly, how have you instrumented that to use daily/weekly and everybody can use in your team to evaluate?

Siddharth: One is about tasks to do based on their impact and time required to finish them. So if a task has a high impact and requires low time, they quite definitely do it. If it has a low impact but low time required, think about doing it. If it has a high impact and will take a long time to do it, get a team or structure a program around it. And if it has low impact and will take a high bandwidth do it, don’t do it. So that's one of the processes I use to eliminate the low-impact tasks out of say 45 tasks by bucketing them like this. Then there is another about the focus of marketing, growth, and GTM. For example, if you are entering a new market, based on the technological maturity, what should a company do or decide or hire or spend for the first 0 to $500K vs $500 to 5Mn vs $5Mn+? Furthermore, we also look at small, medium, and enterprise play here. Next, this framework will also include what role marketing, product, sales, or customer success play in 0 to $500K vs others. So when we are looking at a new geo, we know we need to hire a new person for a particular process. Another framework or rather first, second, and third-order of thinking- say if tomorrow we are planning to grow by 5x in the next two years that directly impact the number of customers we bring in. It looks good from a top-line level but is your finance team capable of handling those many contracts, invoices, and billing? The third order of impact- do you have enough space or folks in your company to onboard these new customers? So every decision that we take follows that order of impact including which teams are affected, and what we are doing to achieve it. For example, tomorrow the Customer Success team wants to upsell. First thing, is a customer success person the best person to handle it or should we move to an Account Executive? So can they pitch and get more out of them? So what is the impact here? Who is responsible, authorized, influencer, or a contributor?

Moderator: The new geography framework, that's the first time I've heard that. Based on the analytic systems that you use today and the challenges that come to your mind, what would be like a wishlist of things that would be either super impactful or are super painful and are not very elegant or not solved completely?

Siddharth: I’ll tell you what I dream about every night. So I want somebody to take my paid data from at least the major publications like Google, LinkedIn, Facebook, etc., and present it to me with the combination of my marketing automation and sales data to report and show CAC or LTV. I don't want to go to siloed databases. I don't want to do vlookups. If a tool can do that for me without me manually entering all the data into a particular tool that works in the backend and gives me data. That is why I probably fell in love with Tableau because once you set it up properly and pull out the right data from API, it is fantastic. But then the problem with Tableau is that it requires a developer. That's an extra resource to have in the company. And at an early or startup stage it is like- do we really want to do that or do you just want to live with inefficiencies in the world? (laughs)

Moderator: So just to again visualize this- you log into a dashboard. It has month level cohort CAC and LTC, and now you want to filter it down by geography, industry, or company size for a particular set. 

Siddharth: Right. For example, if I want to know the sweet spot today, I have to go back to all my marketing and sales data from my sales and marketing automation engine, then I have to ads platform to get my paid data, and then I have to go look at GA data as well. For example from this country, what was the number of visits or what is the sweet spot for that country? Am I spending too much on my paid campaigns? Am I spending too much on my social campaigns? For example, if any of my campaigns guzzles a lot of money and if that can be forewarned that would be super helpful for me because then I can set a  particular cost per opportunity or a cost per lead. So if I set a limit of $X for cost per lead, then I know which campaigns across my network are about to fail. And then I can either look at whether the copy can be improved or the campaign is not targeted right. Actually, my action items become faster and I am more action oriented there. So when you set a particular goal and run with that for 3-6 months, you need heavy data around it to ensure you are doing the right thing.

Audience Question: How much time does you or your team spend on analytics and is anyone from your team tasked with doing all the analytics work?

Siddharth: Me. I don't want my team to break away from their work. At Verloop, we have a person working on the reports with me. I can throw some thoughts out at this person and then the person verifies it for me using reports. But all the analytics ownership is on me. And I spend 50% of my time on analytics and reporting. I can put a bet on it that my entire career has been impacted positively by analytics. And there are a lot of lazy marketers out there who don’t look at analytics. So any marketer who looks at the analytics becomes 2x better that day. Even for a marketing lead owning SDRs, look at the number of calls made or the number of emails or even kind of calls made. Even if you do like a fraction of it, it's fine. The more and more you look at analytics, the better you’ll become at predicting what's going to happen. 

Audience Question: Do you look at our return on ad spends (ROAS) and what is the right range for it?

Siddharth: There is no right range in marketing. You can burn as much money as you want. But in my experience, every marketer should love their finance folks. Get a buy-in from finance. Early-stage startups have 1:1 ROAS. It's not optimal but you're optimizing for growth and not for efficiency. The moment you are like a series B or C and you have some serious money, look at optimization because then the more and more money you end up spending you cannot grow in terms of your trajectory. In the early stage till you raise series B or C, invest in branding and paid generation. When you reach series B or C, your brand and your time in the market should have grown, and your sales motion should also be in play which would then cause your brand and organic to increase and so would your LTV or ROAS.

Moderator: So you are saying that you can spend up to LTV to acquire a customer?

Siddharth: You should because you are optimizing for growth and not for efficiency. Marketing efficiency comes later as far as I've seen. Because, if you're optimized for efficiency first, you don't know how much you can grow. So you will burn money down sometime, you would fail, you would have fallen flat on your face and that is still absolutely fine because then you know what is your limit on spending money for your particular product and your particular market.

 Audience Question: How do you align sales and marketing in the right way where sales and marketing clash over lead volume or quality? 

Siddharth: So this is part culture and part ownership. Culture is where your marketing team and not only the marketing head trust the sales team that they are converting better and everyone on the sales side trust marketing’s campaigns, retargeting, etc. Also, a certain percentage of revenue is owned by both teams. While the salesperson closes it, you do have funnel stages- someone qualifies a lead as an MQL, somebody else qualifies it as a SQL and then you have an opportunity to be created out of it. Your analytics points out in say a weekly or fortnightly manner where the breakages are. This is not something that you get to know at the end of a quarter. From a data perspective, once you are clear on the definitions and SOPs of MQLs, SQLs, and opportunities, you can deep dive into metrics like emails, calls, etc to understand what is breaking. It is also a human tendency- a salesperson might prioritize his/her qualified lead over marketing qualified lead. But if you trust the other person to do their job efficiently then there shouldn’t be a problem. Another thing is if at an organizational level, you differentiate marketing leads from sales leads, then the teams underneath do it as well and the sales team won’t treat both leads the same.

Moderator:  Is there a shared dashboard that both you and the sales team view?

Siddharth: Yes, we review all stages right from all the leads, to qualification, to demo to proposal and we see where it is breaking at region and all other levels.

Audience Question: Do you feel constrained by the time to look into data or you think you're able to look into it?

 Siddharth: I don't think that time is a constraint. If you have the data, you should be able to look into it, provided it does not take a gazillion hours to solve for a single insight if that is not going to impact a lot. 

Moderator: How do you think data can help you improve decision-making? So one of the use cases is budgeting. And I wanted to get your thoughts on two other use cases that we've got from other marketers for the same dream question. One of the use cases that we have got is- marketers keep making changes to campaigns, landing pages, and content pages. Can someone automatically track that so I don't have to put it up in an excel sheet saying on this day, I made this change. And then just overlay that against my metric to say your visitors increased. What did I do right? And so it's sort of tracking. On this day this milestone happened and therefore this impact happened.

Siddharth: That is more like an event tracking feature. From a startup perspective, it is not going to create that much of an impact to keep a track of these changes/events as compared to say Freshworks. But budgeting would be really helpful for me because I can plan my quarterly goals where a database can back into my hypothesis e.g. if I invest 20% more in LinkedIn, I might get 10% more leads. Today, the planning budgeting happens on MS Excel where I manually crunch data from the ads platform, CRM, and marketing automation platform to map monthly budgets with monthly performance metrics- leads, opportunities, etc.

Audience Question: Last question, for how long do you store the historical data? 

Siddharth: 18 months because in a B2B the larger deals take around 12 to 13 months. At a startup, you keep making so many changes and beyond that, I wouldn’t know what changes were made. 

Moderator: That’s all questions we had. Thanks, Siddharth. We are super grateful that you took time out.

Siddharth: Hope it was helpful. Thanks!