Paid Search Management For SaaS Start-ups - An Interview With Soren Ryherd, CEO, Working Planet

soren_headshot.jpg

By Geoff Roberts

Soren Ryherd is Co-Founder and President of Working Planet, a paid search agency located in Providence, Rhode Island. The company is focused on “putting the math behind advertising for data-driven, profitable results.” 

Before we get into the interview, a bit of background - I have hired Soren’s team at Working Planet multiple times. I am admittedly biased, but I’d put it this way - Soren’s company is not just the best paid search agency that I’ve worked with other the years - they are the best agency of any sort that I’ve worked with.

All of that said, I am constantly getting questions around paid search management for start-ups and I continue to see significant challenges at this end of the market. Even further up market and in my past engagements with Working Planet, there are a number of challenges related to paid search management that I’ve felt first hand and still don’t see a great solution to. My intention with this interview is to ask Soren the toughest questions that I’ve had to grapple with as a marketer responsible for paid search performance, and to specifically get his take on how early stage start-ups can set themselves up for success with paid search.

Geoff Roberts: OK Soren, let’s do this. I’ve been thinking about these questions for a while now… I hope you’re ready. Why don’t you start though by giving us a brief background on the work you do at Working Planet.

Soren Ryherd: We build profitable customer acquisition programs using paid digital advertising. I say profitable as we are using our client's’ financial data to determine what to pay, or often not pay, in the digital advertising platforms we manage. Basically we’re doing the math behind the advertising and using that information to remove risk and increase return.

Geoff: Someone once told me... “Soren used to be a rocket scientist. Now he manages PPC campaigns.” Is there any truth to that? And if so, how did you make that transition?

Soren: My Co-founder and I met in graduate school where we were creating algorithms for processing satellite imagery. At that time I had a NASA Graduate Research Fellowship through Goddard Space Flight Center, so I spent a little time there, but they did not let me anywhere near the rockets! After graduate school we worked at separate start-ups and definitely caught the start-up bug. In 2003 we realized that CMOs were really struggling with the auction-based nature of media buying in Search and we knew that was a math problem we could solve in a way that would tie directly to their business success. We kicked the idea around and launched Working Planet two weeks later.

Geoff: What was your familiarity level with Google Adwords at that time?

Soren: This was 2003, and I was working as Head of Business Development for a web engineering company in Boston. We were building websites that “do stuff,” and a lot of our work with tech start-ups was in support of their advertising campaigns. I knew that Adwords was an auction, but I’d never run an Adwords campaign. But people were throwing their hands up, they had no idea what to do with this idea of an auction. And I said “This seems pretty straightforward, you pay what it’s worth.” We knew how to put the math behind this, and we knew we could learn everything else. So we basically just begged some friends to let us run their Adwords campaigns out of the gate, landed some clients, and started doing that.

Geoff: Let’s discuss paid search management for start-ups specifically. Most good paid search agencies have a minimum monthly ad spend that early stage start-ups simply can’t afford. At Working Planet that’s a minimum spend of about $20,000 per month. For a bootstrapped start-up that’s just starting to run its first customer acquisition programs, what options do they really have when it comes to responsibly managing their initial paid search programs?

Soren: It really is a tough question. It is hard to afford the hours to dedicate to a small campaign, but you need to invest the time to build something that can scale.  I would say that the two pieces of advice I most often give to startups are, first, get your data collection nailed down before you start spending money. Good tracking is everything, but I often see startups rush to “try” advertising before they have the tools in place to measure performance. Secondly, know your business KPIs and use those as a lens for evaluating marketing performance. There is really no marketing metric that is good or bad without the context of financial performance. 

Good tracking is everything, but I often see startups rush to “try” advertising before they have the tools in place to measure performance.
— Soren Ryherd, CEO Working Planet

Geoff: How should start-ups begin to think about the tools they’ll need in place to measure paid search performance? What’s a good starting place; how can they identify the tools they might need?

Soren: The first thing you need is some kind of conversion tracking. Everyone’s go-to there is Google Analytics, which has gotten better but which is problematic for us because it’s just aggregate data, it’s very difficult to get per user data out of it. So then people start looking at Conversion Ruler, which is our choice, or Mixpanel. People are just rushing to try advertising, people who aren’t used to thinking about performance based advertising often literally think that all is involved in digital marketing is just launching an ad, and everything happens automatically after that. So people start spending money with nothing in place, they’ll just say “let’s create this ad and target this audience and see what happens.” That’s usually just a waste of money.

Geoff: I personally see a huge gap in the market when it comes to paid search management for start-ups. The good agencies are out of reach price wise. Overseas agencies or freelancers with a lower minimum monthly spend tend to be pretenders rather than contenders. Software tools like Wordstream that automate the management of your spending are a step in the right direction, but tend to optimize for impressions and clicks more so than profitability. Even asking a smart, early employee to manage paid search spending is often problematic - if it’s not something that the person has done extensively in the past, it’s still so easy to burn though budget with these channels. I’m wondering aloud what the best solution for this end of the market might look like…

Soren: You are absolutely right, and most of the software tools out there don’t do much more than what you can do in AdWords or Facebook and often very expensive. What they are trying to solve is to ease management for people who don’t know the networks well, but they have dropped the view of financial performance in doing so. There are a few exceptions. Kissmetrics, for example, has done a great job of pushing financial data into the measurement tools, but that is just on the evaluation side. Meaning Kissmetrics doesn’t manage your bids or your ad campaigns for you, it just provides you the data that can give you the insight to do a good job in managing the campaigns.

In the end though, a startup really looking to scale should invest in digital optimization experience as early as possible. For most of our clients we are driving 50-95% of their customer acquisition and are arguably the most important factor in their success outside of product development.

Geoff: 50%-95%! I’m sure some people will balk at this, feeling that this is being overly reliant on paid advertising or that your clients’ marketing channels aren’t diversified enough. What would your response to that be?

Soren: The reason why this happens is we’re the safe haven. The financial optimization of the campaigns removes the risk from campaign management, making it very safe to invest in marketing. The financial outcome from that investment is very clear, very predictable. So when we can provide that level of reduction of risk and predictability, we get bigger budgets. It becomes a very safe place to invest in marketing compared to almost anything else they could do. 

We’re managing across all paid digital. If someone had 95% of their acquisition in search, then the concern that you have of all of your eggs in one basket is very warranted. The diversification within digital is very important. I think what we’re also seeing is it’s easier and faster to execute with paid advertising, and to optimize to a financial outcome, than it is to with say content marketing, or SEO, or PR. It doesn’t mean companies shouldn’t do those things, they absolutely should, but it’s easier, safer, and more predictable when you look at paid digital. And that’s why we tend to end up with the lion’s share of the customer acquisition. 

There’s also just more in that bucket than there used to be. There are channels that are moving into the bigger digital bucket now that just weren’t part of digital five years ago.

It’s easier and faster to execute with paid advertising, and to optimize to a financial outcome, than it is to with say content marketing, or SEO, or PR.
— Soren Ryherd, CEO Working Planet

Geoff: Ongoing paid search management aside, it’s pretty common for start-ups to want to use Google Adwords or other paid channels to find validation for or initial traction for their product. The instantaneous, real world feedback that you can get from paid search is hugely valuable to start-ups in that way. What best practices, strategies, or guidance would you give to start-ups that want to use Google Adwords as a means to validate a market or idea?

Soren: First, unless you are testing messaging, DON’T test for “traction” until you have built your payment functionality. We’ve seen many companies “prove” their business model through free sign-ups only to have crickets chirp the second they ask for money.  There are many valid ways to use digital advertising to help in the early validation stages of a start-up, including finding real customers to speak to in the MVP development process, determining key selling propositions to early users, and seeding test users into applications while in development. I would say the key is to be really explicit in what you want. The absolute worst goal is to “get exposure.” Getting exposure means that there are assumptions about the value of that exposure that you haven’t surfaced. Is this a euphemism for driving product demand? For getting investor attention? That doesn’t happen by magic and the most dangerous thing is to think that delivering ad impressions is in any way a goal.

Geoff: I recently installed Adroll on Outseta’s website so that I can start building an audience of visitors to run remarketing campaigns to. What other easy items represent low hanging fruit and are things start-ups can do now to set themselves up for success with paid online customer acquisition programs?

Soren: AdRoll retargeting is a very good move, but only leverages people being driven to your site from other initiatives. The first thing we would suggest is getting ads on your brand out there in Google and Bing (yes, really, Bing). Use your core value proposition and say something different than in your organic listing. Secondly, if you already have a customer list, there are many targeting options now for creating lookalike audiences which tend to work well. Another low-hanging opportunity is ads on your competitor brands in Google, although I will say that many startups are often reluctant to do this as they (usually mistakenly) think that by not doing this they are staying off the radar of the big guys. What we find is that if you are new, leveraging the brand equity of the existing big players works to your advantage.

Geoff: Let’s get into one of the biggest challenges I consistently see start-ups face with paid search programs - expectation setting with the C-suite. It’s very easy for paid search programs to be seen as a major expense - they are often one the of the biggest line items in a marketing budget, and at the outset campaigns almost never yield the desired results. Making these programs work is a process that takes commitment and quite a bit of iteration. How do you go about setting appropriate expectations with the C-suite in terms of what it will take find the value of paid search programs to their business? If you could get Founders, CFOs, etc to all sign an agreement outlining what expectations they should have when you start working together, what would it say specifically?

Soren: One of our biggest goals with any new client is to move digital advertising out of the cost column and to be seen as a profit center. But to do that, we have to connect the dots so that the C-suite knows that X dollars in yields Y dollars in profit - and they have a say in what X and Y look like. We think that the CFO should absolutely be in the conversation, and (I guess not too surprisingly) CFOs tend to be our best friends and internal champions.  

One of the most important things to making profit-based campaigns work is having good targets, in terms of cost of acquisition. This is something we spend a lot of time on with our clients. If you just accept “Get us $100 leads” as a goal, you are never going to be well optimized, as all leads are not equal in value and should not be paid for equally, and, frankly, someone probably just made up the $100 number in the first place. Better targets are those that will truly raise profitability.

Geoff: Do you have any advice on how SaaS businesses should go about identifying a good CAC target?

Soren: Every company is going to have a different comfort level on cost of acquisition relative to predicted lifetime value. Some of that comfort is going to be based on their funding, and just available cash that they can put into marketing. Some of it is going to be how much data do they have to show the predictability of what they think lifetime value might look like. LTV has become a little less part of the discussion, surprisingly, as we look more at what are the predictable ways to raise MRR and manage cash flow effectively. So months to break even tends to come more into play relative to those things. 

The higher value the customer and the more embedded the technology is in the client’s company, we tend to see bigger multiples in terms of months to break even - you might see something like one year or 18 months. When churn is a question, when LTV is a question, or if it’s just easier for companies to churn out of a technology, we’re probably going to see a lower target in terms of months to break even.

The great thing about digital is it’s not difficult to change your acquisition targets. You shouldn’t do it all the time, but you should be influenced by how the data matures over time. If you’re a SaaS client and you don’t have a churn issue, if you’re targeting 1-2, or 3 months to break even as your cost of acquisition target, you’re probably undervaluing the market. By a lot. Your LTV is so high compared to what you are willing to invest in that customer, that you’re really ceding opportunities to competitors at that point. 

Geoff: So when expectations with the C-suite aren’t set appropriately, I see two things tend to happen. First, too little is actually spent month to month, meaning it takes a really long time to find which channels, bids, and campaigns can yield the results you are looking for. The company doesn’t see a big impact on their business after a few months, so they give up. The second is the company spends enough money, but doesn’t have the patience it takes to optimize the campaigns over time. After a couple of months of significant spending not yielding the desired results, the company throws up its hands and says “paid search just doesn’t work in our industry.” What’s your viewpoint - can paid search work in just about every industry?

Soren: Yes, but not always at scale, and not always immediately. All digital, and search is a great example, has volume that is top-loaded in the auction based on ad position. Every advertiser wants the volume and therefore top position. But often the math just doesn’t work to have your ad show up there, so you are losing money in volume. Optimization means learning over time how to value the audience appropriately based on the value they create. A lot of the time we are learning when not to spend money, or what needs to change in terms of conversion, close rates, and customer value in order to afford the advertising. The good news is that data and complexity help as we can find the audiences that work while attacking “the math” through testing and optimization to achieve scale.

Geoff: OK, so Working Planet talks a lot about “putting the math behind advertising,” about profitability, about “closed loop reporting.” Let’s talk about what this actually means for a SaaS company when it’s working properly. For me, the marketer, the coolest part of seeing this come to fruition was that each month our paid search spending would be a fixed cost - say we spent $20,000 on paid search programs in July. In each subsequent month when I looked at my month end report from Working Planet, I could see how much revenue that $20,000 in spending had actually generated. Because it’s SaaS the revenue realized number would grow each month, I could see exactly when we reached break even on any month’s paid search spend, and I could see the profitability of any month’s spending continue to grow thereafter. Of course as customers churned, you’d fold that in as well and would see that also impact the actual revenue that was realized. It’s incredibly powerful to have that level of understanding of the relationship between your spending and your profits. But let’s talk through what it takes to get there - proper lead source attribution, billing and cancellation information synced with CRM records, that sort of thing. Talk us through the integration and technical work that your clients need to understand and commit to to truly “close the loop.”   

Soren: We use three buckets of data that, in combination, give us the ability to make good decisions on buying digital media. First, we have cost and targeting data from the ad networks like Google, Bing, Facebook, Twitter, LinkedIn, DSPs, etc. Second, we have multi-touch per-user tracking on our client’s site to measure engagement, and; Third, we have customer value data from our client’s CRM, Finance, or other internal systems. We work the “loop” both backwards from the customer value data to create our models, and forwards through the user engagement path to monitor performance by granular segment.

The client-side systems and data are always the wild card.  While all we really need is the common “join” to marry the back-end data to the marketing data, we always find significant data gaps, multiple “systems of truth,” or missing revenue when we dig in. Inconsistent use of a CRM by the sales team is also a huge issue.

For SaaS clients, we are typically looking at a months-to-breakeven target that is related to growth of MRR. In some ways SaaS clients are easiest in that there is a built-in recognition that the investment in marketing does not occur in the same time period as revenue realization. This point is lost on a lot of companies that only take a “cash flow” view of in-month expenditures and revenues. This can cause real problems in evaluating performance so we try hard to introduce a cohort-based view of that investment and how it pays off over time.

DON’T test for “traction” until you have built your payment functionality.
— Soren Ryherd, CEO Working Planet

Geoff: Let’s talk about tracking scripts and pixels - admittedly these have become sort of dirty words to me. In order to deliver on closed loop reporting, you need to add tracking scripts or pixels all over the place - website pages, lead capture forms, etc. Depending on the actual paid search network, or remarketing network, or social channel you are using, there seems to almost always be another tracking script that needs to be added somewhere. And all of this tracking work needs to happen consistently across different device types and different browsers. 

I’ve enlisted the help of plenty of very technical people who have sort of poo-pooed how difficult this work could possibly be, only to find that it’s almost always more… finicky... than they initially expected. I’d go so far as to say that at times I felt like we even spent more time working on tracking scripts than on the campaigns themselves. Yes, there are tools out there like Google Tag Manager but frankly they don’t solve this problem. How do you think about this and address the pain that customers, including myself, have felt in this area?

Soren: Well, first, I would not knock GTM as it solves a lot of the pain that we experience with scripts and pixels not being viewed as supported technology by our clients’ web teams. If you want data-driven optimization your only source is in fully utilizing all the scripts and pixels from all the networks, and not just your primary tracking tool.

A bigger issue is, in my opinion, in the limitations of tracking technology. We are a very technical data-driven shop, and our biggest “data challenge” is getting clients to buy into cross-channel behaviors. For example, 95% of engagement from a Twitter campaign will come from people who never click on the Twitter ad. Some clients still believe that if the click didn’t happen, the channel had no influence (and, conversely, “How do we buy more of this “Brand” traffic? That performs great!”).  This is such an issue that we are now trying to report 100% of paid and non-paid engagement in all our reporting plus explicit testing to validate out-of-channel lift.

Geoff: Can you give me an example of how you validate out of channel lift with paid advertising campaigns?

Soren: With a start-up client that we’re working with right now, we knew we were getting significant out of channel lift from Facebook campaigns. It was very difficult to convince them of how much that might be - because it never involved a click on an ad it couldn’t be tied to Facebook specifically. 

So we did some geographic tests where we used Texas as our test area, and Florida and Georgia as the control. We dramatically escalated Facebook advertising in Texas, and what we found was there was a 90%-100% lift above what we directly tracked to Facebook showing up in their brand and no-referrer traffic, meaning 45%-50% of all value from Facebook was being created without a click on an ad.

This allowed us to create a really good model for how we should think about cost of acquisition targets for tracked Facebook relative to the whole campaign. We then did a confirmation test using Colorado and Utah, and we found the exact same behavior. We were able to validate for the client that it wasn’t just 60% of their acquisition that was from the paid campaigns, but actually over 80% of their entire business that was being driven by their paid campaigns when you factored in the out of channel lift.

When you can nail down the math on this and predictably show that an outcome is going to exist, and you recalibrate that on an ongoing basis, what we find is people are going to give you an unlimited budget. One that’s limited only by their own limitations as an organization, which might be financial in which case a capped budget is appropriate. Otherwise it’s going to be restricted only by cost of acquisition targets relative to the competitiveness of the market.

Geoff: Beyond the mechanics typically associated with paid search management - keyword selection, bid optimization - talk to me about what else goes into optimizing paid search campaigns over time. What I’m getting at here is the client usually has to play an active role to truly allow your company to do it’s best work whether it’s customer research, feedback on new messaging directions to test, conversion optimization, etc. 

Soren: The “math” we are solving for involves the entire engagement path, so we have always advised on page testing as one of the key additional areas where we can help scale. Last fall we launched a Conversion Design Service for a small monthly retainer that allows us to now build and deploy landing page tests as well, accelerating learning and optimization for clients that don’t have those resources in-house. Our clients are very involved in this as well, addressing other core factors in the equation through nurturing programs, lead response times, proper lead scoring, use of SDRs, and retention programs. Every gain in the entire acquisition/LTV chain makes it easier to scale the programs.

Geoff: If there was an idea or two that you could beat into the head of start-up Founders when it comes to paid search or digital advertising more broadly, what would it be?

Soren: This is a process of ongoing, continual improvement that will never end. That’s it, really. I guess if I were really hammering it home it is that what you are improving on is the ability to generate revenues and profits. 

Geoff: Let’s end by tying this all back to Outseta for a moment. We are a platform play that competes in an ultra competitive market, across extremely crowded categories like CRM, email marketing, customer support, and subscription billing software. On top of that, the vast majority of our competitors in these categories are established or very well funded businesses. How can we possibly compete in paid search? How would you think about the role that paid search should play in our business, or the strategy/approach to paid search that might make the most sense for us?

Soren: Search is a media that serves up solutions to stated problems, so for search, you need to look at the many problems you are solving that people are searching on. However, many start-ups solve problems that exist, but are unstated or unsearched on because people can’t envision the solution exists (imagine selling a self-driving car before people knew this was technically feasible). In those cases, you need to create that awareness with media outside of search, but there are more options for that everyday in Paid Social, Display, Programmatic, Video and other platforms. For Outseta, like many SaaS start-ups, it will be a process of testing many varied potential audiences against their actual engagement in the sales process.

Geoff: That’s a wrap. Thanks for your time, Soren!