FINE-TUNING FINTECH How we took a B2C financial services provider from the automated bid “death spiral” to unprecedented growth

CLIENT CHALLENGE SOLUTION ◊ Financial services provider ◊ Automated bidding had put ◊ Hands-on, detail-oriented specializing in online bill- them into the “death spiral” account management paying ◊ Impressions, clicks, and ◊ Adjusting keywords for more ◊ 45 billers nationwide in the conversions were all down by targeted traffic energy industry at least 33% ◊ Improving ad copy and quality ◊ Revenue was down 42% scores for cumulative savings ◊ of $630,000 OUTCOMES

COST PER CONVERSION 78% CONVERSION RATE RETURN ON  AD SPEND 417% 431% “In just the first year, (un)Common Logic fully crafted and optimized our program— not just our paid search account, our entire paid search program. I truly believe they’ve left no stone unturned in their pursuit of high returns. (un)Common Logic fully has diligently and thoughtfully positioned us in the PPC space better than any other provider has or, in my opinion, could have done.” —VP of Marketing A financial services provider specializing in online bill paying approached us for help with their paid search advertising. They had been managing their account using automated bid management tools, but the results were alarming. They were caught in the “death spiral” of paid search. Automatic bidding had lowered their bids, and thus their impressions, clicks, and conversions. In the past three months, they had seen: ◊ 33% drop in impressions ◊ 38% drop in conversions ◊ 35% drop in clicks ◊ 42% drop in revenue

THE DANGER OF AUTOMATED BID MANAGEMENT Automated tools can be very appealing. Some of them are built into a paid search platform, making TOOLS CAN GIVE THE ILLUSION them extremely convenient, especially for an in-house team that might not have much expertise in the de- THAT AN ACCOUNT IS CONSTANTLY tails of account management or the budget to screen tools and purchase one or more. BEING FINE-TUNED FOR OPTIMAL Tools can give the illusion of oversight and efficiency, PERFORMANCE. HOWEVER, making it look like an account is constantly being ad- justed and fine-tuned for optimal performance. How- “OPTIMAL PERFORMANCE” CAN ever, “optimal performance” can vary depending on the VARY DEPENDING ON PERSPECTIVE... perspective from which it is evaluated.

◊ It’s in a client company’s best interest to optimize Without strong, detail-oriented oversight to harness performance toward bottom-line metrics such as bid adjustment tools, they could make changes that conversions, revenue, and cost per conversion completely undercut bottom-line performance. ◊ Without initial adjustment and constant manage- We don’t make a policy of avoiding tools; on the con- ment, a tool could optimize based on non-reve- trary, we use many of them to optimize our clients’ nue-related metrics such as average ad position, paid search activities. However, we vet each tool we total spend (rather than cost per conversion), or use very carefully, and only use it to the extent that it click-through rate makes us more efficient in carrying out the strategies and tactics we’ve designed for our clients.

(un)Common Logic 2 In this client’s case, over-reliance on automated bid ◊ Target costs per conversions had been lowered management tools had caused these problems: so many times that for a significant number of ◊ High-volume keywords had been paused, keywords, bidding high enough to get an ad severely limiting traffic and thus conversion visible on a results page was opportunities essentially impossible ◊ Because keywords couldn’t be reenabled Automated bid management, intended to boost automatically, these high-volume terms stayed efficiency, had instead torpedoed performance and paused, exacerbating the effects of lowered traffic harmed the company’s bottom line. Fortunately, they contacted us before any more damage could be done.

PLAN FOR IMPROVEMENT Ordinarily, our strategic plans have three phases: However, this client’s accounts were complex and in- ◊ Phase 1 delivers short-term fixes with visible cluded dozens of individual billers, each with different results in less than 1-2 months data-gathering methods. So we folded Phases 1 and 2 together, improving and rebuilding campaigns on a ◊ Phase 2 rebuilds the client’s paid search account per-biller basis over the next 8 months. during months 3-6 More than five years after the start of our engage- ◊ Phase 3 expands the reach, power, and capacity ment, Phase 3 is still ongoing because optimization of the client’s paid search by testing new chan- never truly stops. nels, including remarketing and mobile

IMMEDIATE FIXES During the first phase of our engagement, we used stabilize the account and making simple improve- a triage approach: fixing what needed to be fixed to ments that yielded significant returns.

STABILIZE THE ACCOUNT: MANUAL BIDDING & OPTIMIZATION The first, and most important, thing we did was to algorithm and gave it to our paid search experts, who change all campaigns from automated bid manage- combined their expertise with our mission to improve ment to manual bid management and optimization. our clients’ returns and revenue – not those of the By doing this, we took control from the platform’s platform. DELIVER QUICK RESULTS: KEYWORDS The client’s keyword approach was counterproductive ◊ Many ad groups had duplicate keywords to attracting high-value searchers: ◊ Not enough negative keywords were used to ◊ High-volume, relevant keywords had been paused weed out non-relevant searches or deleted By fixing all these keyword issues, we refocused the ◊ Some keywords strongly related to business goals account toward better searcher intent to increase had been marked as negative traffic, conversions, and revenue.

(un)Common Logic 3 PHASE 1 & 2 RESULTS

Our engagement with the company Phase 1 & 2: Conversions & Conversion Rate began in May 2012. By December, 35,000 35% we delivered these improvements 30,000 30% compared to the first 4 months of 25,000 25% 2012: 20,000 20% ◊ 181% increase in conversions 15,000 15%

◊ 275% increase in conversion rate 10,000 10% (CVR) 5,000 5%

◊ 68% reduction in cost per 0 0% conversion APR-12 MAY-12 JUN-12 JUL-12 AUG-12 SEP-12 OCT-12 NOV-12 DEC-12 CONVERSIONS CONVERSION RATE ◊ 10% reduction in average weekly spend Phase 1 & 2: Revenue & ROAS ◊ 193% increase in average weekly $120,000 500% 450% revenue $100,000 400% 350% ◊ 224% increase in return on $80,000 advertising spend (ROAS) 300% $60,000 250% 200% $40,000 150% 100% $20,000 50% $0 0% APR-12 MAY-12 JUN-12 JUL-12 AUG-12 SEP-12 OCT-12 NOV-12 DEC-12

REVENUE RETURN ON ADVERTISING SPEND

PHASE 3: TESTING NEW CHANNELS After we had fully rebuilt the client’s Google account As we do with every channel, we customized: and demonstrated consistently improving perfor- ◊ Ad copy mance, the client gave us additional budget to add new channels. We began by adding Bing as a paid ◊ Keyword choice search channel in the second year, then added Yahoo ◊ Bid settings and strategies Gemini in the third year.

(un)Common Logic 4 BING RESULTS Conversion Volume After Adding Bing 180,000 Because of our attention to detail when we 160,000 added Bing and set up its account and cam- 140,000 120,000 paigns, we saw gradual results that made a big 100,000 difference over time. Over the next 6 quarters: 80,000 60,000 ◊ Conversion volume had more than doubled, 40,000 reaching nearly 160,000 per quarter 20,000 - Q3 13 Q4 13 Q1 14 Q2 14 Q3 14 Q4 14 ◊ By the end of Q4 2014, Bing was con- GOOGLE BING TOTAL tributing almost as many conversions as Google: 21,046 to Google’s 22,929 Revenue:Conversions Ratio, Bing vs Google 1.08 Bing has delivered consistently high-dollar 1.06 conversions; except for one period in mid- 1.04 2014, its revenue-to-conversion ratio (the 1.02

Value Ratio) has been 1.00 or greater. 1.00

0.98

0.96

0.94 Q3 13 Q4 13 Q1 14 Q2 14 Q3 14 Q4 14

GOOGLE BING

Revenue:Conversions Ratios, Yahoo vs. Bing vs. Google YAHOO RESULTS 1.10 Adding Yahoo in January 2015 further

1.05 expanded our client’s reach. Like Bing, Yahoo has more female users than 1.00 male. As more than 65% of household 0.95 bills are paid by women, Yahoo was a logical choice for our third channel. 0.90 Yahoo Gemini hasn’t historically add- 0.85 Q1 15 Q2 15 Q3 15 Q4 15 Q1 16 Q2 16 Q3 16 Q4 16 Q1 17 Q2 17 Q3 17 ed a great deal of conversion volume. GOOGLE BING YAHOO However, its cost per conversion is comparable to both Google and Bing, Conversion Volume After Adding Yahoo Gemini and its cost per click is much lower. 550,000 Additionally, like Bing, Yahoo Gemini’s 450,000 relative value per conversion has been

350,000 higher than .99 for 18 months now.

250,000 Google is still the channel of choice for volume, but these other channels reach 150,000 the client’s target audiences more effi- 50,000 ciently and deliver proportionally more

(50,000) Q1 15 Q2 15 Q3 15 Q4 15 Q1 16 Q2 16 Q3 16 Q4 16 Q1 17 Q2 17 Q3 17 revenue.

GOOGLE BING YAHOO TOTAL

(un)Common Logic 5 PHASE 3: MOBILE

We determined that mobile would probably be Mobile Performance: Conversions a strong channel for our client because of its 200,000 35% existing presence in their industry. For instance, 30% 150,000 according to research conducted by the Pew 25% Charitable Trusts, by 2015, one-third of U.S. 20% 100,000 adults had used a mobile device to pay a bill. 15% 10% We introduced Google Calls and Bing Calls 50,000 5% in Q2 of 2014, and both channels performed 0 0% consistently well. The client increased their Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 14 14 14 15 15 15 15 16 16 16 16 17 17 17 investment in mobile in 2017, leading to even MOBILE CONVERSIONS % OF TOTAL CONVERSIONS greater gains. By the end of Q3 2017, mobile was responsible for nearly 30% of total revenue Mobile Performance: Revenue and total conversions. $500,000 35%

30% Revenue:Conversions Ratio, Mobile vs Desktop $400,000 25% 1.06 $300,000 1.04 20% 1.02 $200,000 15% 1.00 0.98 10% $100,000 0.96 5% 0.94 0.92 $0 0% Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 14 14 14 15 15 15 15 16 16 16 16 17 17 17 14 14 14 15 15 15 15 16 16 16 16 17 17 17 MOBILE RATIO DESKTOP RATIO MOBILE REVENUE % OF TOTAL REVENUE

PHASE 3: QUALITY SCORE

One often-overlooked aspect of traffic, and conversions. However, Quality Cost per click relative paid search is Quality Score, the a higher Quality Score will also Score to 7.0 Quality Score numeric value a search engine result in a lower cost-per-click 10.0 30% lower assigns to a keyword based on bid, as search engines discount how well a client’s ads adhere to CPC bids for scores over 7.0 and 9.0 22% lower best practices. For all three major increase bid costs for scores lower 8.0 13% lower search engines, Quality Score is than 7.0 (ads scoring 7.0 receive 7.0 baseline rated on a scale of 1 to 10 and no discount or inflation). 6.0 17% higher comprised of these criteria: Our client had re-written or re-in- ◊ Relevance to searcher intent troduced ad copy many times in 5.0 40% higher ◊ Expected click-through rate the months prior to our engage- 4.0 75% higher ment with them, so many Quality 3.0 133% higher ◊ experience Scores were reset to 7.0. Thus, the Of course, higher-quality ads will average of all their scores was 7.1. 2.0 250% higher result in more visibility, targeted 1.0 600% higher

(un)Common Logic 6 QUALITY SCORE RESULTS Over the next 5 years, Cumulative Savings from Quality Score Gains we brought their average Quality Score up from 7.1 $600,000 to 9.0. Not only did the high-quality keywords $500,000 deliver more relevant traffic to the client’s site, the $400,000 increase in Quality Score resulted in significant $300,000 savings. $200,000 ◊ By the end of 2017, monthly savings $100,000 averaged around $28,000 $0 16 13 14 15 17 16 12 13 14 15 17 16 12 13 14 15 17 16 12 13 14 15 17 ------FEB ◊ As of Q4 2017, we FEB FEB FEB FEB MAY MAY MAY MAY MAY MAY AUG AUG AUG AUG AUG AUG NOV NOV NOV NOV NOV NOV have saved the client a total of $629,557 in costs per click alone No automated tool could deliver these improvements in Quality Score; that requires human intelligence, expe- rience, and insight, as well as the diligence to keep improving ad copy and landing pages.

OVERALL RESULTS

Not only did we get our client out of the Annual Revenue from Paid Search “death spiral” in less than a year, we built a $5 strong foundation for growth. We improved every metric: conversion volume, conversion $4 rate, cost per conversion, return on advertis- ing spend, and revenue. $3

From the start of our engagement to the end MILLIONS $2 of 2017, we have achieved these results: $1 ◊ 1,495% increase in weekly conversion

volume $0 2012 2013 2014 2015 2016 2017

◊ 417% increase in conversion rate BEFORE US ◊ Cost per conversion reduced by 78% For 2016, we set a goal with the client of earning ◊ Weekly revenue increased by 1,789% while $2.1M in revenue by the end of the year. We exceed- weekly ad spend increased by 258% ed that goal by $134,000. And in 2017, we more ◊ 431% increase in return on advertising spend than doubled it, with a revenue of $4.96M.

(un)Common Logic 7 When our client approached Conversions & Conversion Rate Over Time us, they had entered the death 2.0 40% spiral of automated bid tools: as costs-per-conversion were automatically bid lower and 1.5 30% lower, the company was losing more and more visibility, traffic, 1.0 20%

conversions, and revenue. MILLIONS

We immediately assumed 0.5 10% manual control of their ac- counts and began applying our 0.0 0% expertise, insight, and atten- 2012 2012 AFTER 2013 2014 2015 2016 2017 tion to detail. We adjusted and BEFORE US US optimized their bid strategy, CONVERSIONS CONVERSION RATE and achieved an astonishing turnaround by the end of that Conversions & ROAS Over Time year. 2.0 500% In the next few years, we 400% added more channels based on 1.5 research into the client’s target 300% market, their bill-paying be- 1.0 havior and their search engine

MILLIONS 200% usage habits. We expanded 0.5 into and 100% saw significant results from that channel. 0.0 0% 2012 2012 2013 2014 2015 2016 2017 Throughout our engagement, BEFORE US AFTER US we have increased their ads’ CONVERSIONS RETURN ON AD SPEND average Quality Score so much that they’re now saving almost Revenue & Cost per Conversion Over Time $30,000 a month in bids and $5 $3.50 getting a much better return on their investment in paid search. $3.00 $4 CONVERSION PER COST $2.50

$3 $2.00

$1.50 $2

$1.00 REVENUEIN MILLIONS $1 $0.50

$0 $0.00 2012 2012 2013 2014 2015 2016 2017 BEFORE US AFTER US

REVENUE COST PER CONVERSION

(un)Common Logic 8 TAKEAWAYS The steady growth we’ve produced for our client ◊ Focus on Quality Score: didn’t come from a one-size-fits-all approach or a The higher your ads’ Quality Score, the lower hands-off style of just adjusting some campaign your cost per click. Plus, higher-quality ads attract settings and monitoring the results. Instead, they’re a higher-quality traffic, leading to more conversions natural development from our philosophy regarding and revenue from fewer clicks. A high Quality digital marketing, especially these aspects: Score is key to maximizing your return on adver- tising spend. ◊ Autopilot is not an option: If you let a tool or platform control your search ◊ Always be testing and optimizing: marketing, it’s bound to make decisions about Never settle for “good enough” or “what worked your spend that won’t benefit you. It’s always best before.” Keep seeking ways to improve perfor- to have human intelligence in charge. mance, remove friction from the buying process, and introduce beneficial innovations.

(un)Common Logic solves the hard problems in digital marketing by using data to uncover surprising details, then using human intelligence to leverage that information for uncommon results. Visit www.uncommonlogic.com or reach us directly at contactus@ uncommonlogic.com.

9