PPC Chat Streamcap – Best Practices for Implementing Tests & New Strategies

This week Luke Alley (@LukeAlley) is guest hosting and helping us sharpen the PPC saw with “Best Practices for Implementing Tests & New Strategies.” The following is the transcribed Streamcap from the live chat:

Q1: How often do you run “tests” you’ve never tried before? i.e. new strategy, new ad formats, etc.

  •  Awesome question, no its not weird, black is the most common cap color, so Google wanted to send you something but not invest. Always-Be-Testing! Whether its running a different ad version, trying out some new kws, or changing a LP element. Larger tests, take some planning, but should be scheduled in when appropriate. Accnts not being tested tend to degrade over time. I am also speaking from the client side – i come up with ideas, and agency executes. As an agency I would try to give the client test ideas so I can control the experiment. Part of a status update. – Chris Kostecki (@chriskos)
    •  True! How about for more advanced strategies or bigger tests (however you define that)? – Luke Alley (@LukeAlley)
      • Promos may trump a big test, otherwise I am trying new targets every week or 2, some flame out quickly others survive. – Chris Kostecki
  • Always! I’m always on the lookout for a new test I haven’t tried on an account. I do recycle though! – Heather Cooan (@HeatherCooan)
  • If the acct is in a good spot and the client is ok with spending, then I say do it as much as you can. – Neil Sorenson (@iNeils)
  • When metric plateaus seem to be reached, issues seem to be occurring or if requested by client. – Nicole Mintiens (@Tregesy)
  •  when I have an idea I put it on TODO list. Normally a low priority item though so depends a lot on time. – Richard Fergie (@RichardFergie)
  • Honestly not as much as I should. I tend to go with tried and true until I hit a wall, then test new stuff. – Aaron Levy (@bigalittlea)
    • Playing it safe is a valid strategy too! I’m just a rebel by nature! – Heather Cooan
  • Anytime Google launches something new… on my own stuff first if I can. – Elizabeth Marsten (@ebkendo)
  • Not as often as I would like for completely new strategies. Consistently testing, yes. Reinventing the wheel each time, no. – Michelle Morgan (@michellemsem)
  • Depends on how crazy and/or hard to implement the new idea is. Also, time can be a challenge. – Melissa Mackey (@Mel66)
    • Absolutely. Time/cost to implement is important – new kwds = cheap. New landing pages = not so much. New business model = !! – Richard Fergie
  • Groundbreaking stuff – unfortunately it’s discussed more than implemented for me. Still, always testing some item in the account. – Francis Shovlin (@fmshovlin)
  • It also depends on the flexibility of the site and the analytics tools used…sometimes it’s a fight. – Heather Cooan
  • Bidding is the junk food of PPC. Easy2analyze+ Easy2change=immediate gratification. Real search marketers focus on testing. – Chris Haleus (@chrishaleua)
  • Time seems to be coming up a lot. How much time SHOULD you devote to testing… in your opinion?
    • A bit more than I do right now. Bet everyone says this. – Richard Fergie
    • This is a difficult question, seems like everything I do is a test. – Heather Cooan
    • There’s never enough time to do all I’d like to do, whether testing or anything else. – Theresa Zook (@I_Marketer)
    • It’s really up to the client. get them on board, because many tests fail. – Chris Kostecki
    • As often as the data/client allows. Never a good idea to be stagnant. Id like to section off part of my time to ideas, testing. – Francis Shovlin
    • As much as is warranted by your current workload. – Nicole Mintiens

Q2: What are your “best practices” for running tests and new strategies in PPC accounts? 

  • Gotta pick one thing, control. Gotta have enough data to decide if it works. Gotta collect data post click! – Heather Cooan
  • Trying to align an equal control vs. experiment. Virtually impossible given time weather etc., but the more aligned, the better. – Chris Kostecki
  • With new strategies I like to run the test over multiple similar accounts and review results separately & combined. – Cassie Allinger (@CassieAllinger)
  • First, know what you’re testing FOR. Second, run detailed reports in advance, for solid data to base change analysis on. – Theresa Zook
  • A/B if possible. Otherwise: establish a baseline, standard deviation of metrics & compare (& hope outside factors arent major). – Francis Shovlin
  • Clearly defined goals and hypotheses is a must…need direction! = Heather Cooan
  • Clearly outlined! What is being tested? Goal? Start/End Dates. Expected result (i.e. How will success be determined?). – Nicole Mintiens
  • I actually recently created a scientific method sheet for testing that has helped make testing more uniform. – Noah Brooks (@noahbrooks)
  • Don’t test 2 many variables at once, Give tests enough time to get results, Use the scientific method, Understand the context. – Chris Haleua
  • A/B (if possible) – 30 day test period with a 15 day check in. And be scientific! – Aaron Levy
    • Yes, 30-days good time frame. – Heather Cooan
    • I think you need to wait that long to know something works but can often tell if it doesn’t a lot quicker. Fail fast! – Richard Fergie
  • Once you think you have the best ad/langing page ever, TEST IT AGAIN ANYWAY and you will learn there is always room to improve. – Chris Haleua
  • Only change one thing at a time, ad copy, landing page tweaks, different landing pages so you can KNOW what works. – Bryant Garvin

Q3: How do you determine which tests you should run for your accounts?

  • By the goal…more traffic, more clicks, more conv, what am I after? More goals…lower CPA, higher ROI, higher margine, lower CPC, more visibility, higher ad position, etc. – Heather Cooan
  • Whatever makes sense for the client and their goals. – Francis Shovlin
  • Tests should always align with the goal(s) – They can also be prioritized based on their relevance to the goal. – Cassie Allinger
  • Depends on what I’m doing already, or what I’ve already tried. (And client goals.) – Theresa Zook
  • Assessing the funnel, is the biggest issue CTR, CVR, or ROI. – Chris Kostecki
  • What is the goal you are shooting to improve? What previous tests have you run? That helps determine what tests next. – Bryant Garvin
  • Prioritize by client and goal (and expected likelyhood of effectiveness sometimes). – Nicole Mintiens
  • For me it’s the account goals…which should always be consistent with overarching marketing/client goals. – Heather Cooan
  • Knowing specifically what we want to improve & developing strategies that can make it happen. Always keep long term goals in mind. – Michelle Morgan
  • With 12 sided dice of course. Let fate decide! – Elizabeth Marsten
    • I prefer 2 D12 for ad copy & 1 D8 Landing pages. – Bryant Garvin
  • Start with the big changes (consistently work) then move on to those where little tweaks can make small % boosts. – Bryant Garvin
  • Like others have said depends on goals of the account. If all is going well good to test if we can improve CTR/CPA/etc. even more. – Luke Alley

Q4: Where do you find inspiration for new tests? Blogs? Twitter?

  • PPCChat of course. – Cassie Allinger
  • Blogs, networking, conferences, twitter chats, new product releases, client strategy meetings, all over the place! – Heather Cooan
  • Outlandish client requests? – Nicole Mintiens
  • Testing blog inspiration via@antfoodz. – Chris Haleua
  • Seriously? Twitter, Blogs, etc but I REALLY try to understand biz & potential customers put myself in their shoes. – Bryant Garvin
  • Competitors and my fav way is regularly revisiting unrelated, competitive niches for ideas in both Search & Display/Content. – James Zolman (@jameszol)
  • New opportunities are found is social, listening to what the audience is saying and reacting to it. & analytics, email efforts. – Chris Kostecki
  • Getting into the psychology behind behavior helps me in a lot of my ideas for testing. – Bryant Garvin
  • Look at new products, forums, twitter and competitors. – Matt Dyer (@mattchstik)
  • Experience with previous clients. If it worked for one it could work for another. – Brett Stevens (@BrettStevens1)

Q5: What’s the most creative and succesful test/strategy you’ve ran?

  • Most creative – integrating CRM data w/GA for segmentation in retargeting. – Heather Cooan
  • Those tend to be LP tests. – Theresa Zook
  • Lips are sealed – my best results are still out performing competition. – Chris Kostecki
  • Seen plenty of wins testing location targeting breaking out outlier locations (good or bad), but not a guaranteed win. – Noah Brooks (@noahbrooks)
  • Most successful – increased brand trust with communication extensions & sitelinks together. Issue w/counterfiters. – Heather Cooan
  • I set up a ‘November Madness’ tourny for my value props to be A/B tested throughout ad copy. Still running so success is TBD. – Eric Farmer (@click_eric)
  • Tested weekend day parting with one of my campaigns. Last vs. Prior 30 lowered costs by 45% increased revenue over 10%. – Brian Gaspar (@BGaspar)
  • Some of best wins are still simple one word, punctuation, etc changes in Ad Copy. – Bryant Garvin
  • Best results have been through simple LP tests: changing headlines, button color/size, etc. Seen 15%+ lift in conv rate. – Francis Shovlin
    • Agree, I tend to see bigger lifts with LP tests than ad copy (when goal is CVR). – Ginny Marvin (@GinnyMarvin)
  • I agree with Bryant. Landing Page Optimization seems to deliver the most noticeable results. – Eric Bryant (@GnosisArts)
  • Symbols and numbers still work wonders. Not entirely creative, but they work for me in a lot of cases. – James Zolman
  •  Propose additional cost and expected outcomes. If implemented tests before, show how those have improved their performance. – Francis Shovlin

Q6: How do you address testing with your clients/boss/whomever?

  • Usually I’ve built trust with an ROI lift in an audit so pitching testing is easier. Gotta build trust first. – Heather Cooan
  • Scope, Duration, Impact, Anticipated results. – Brian Gaspar
  • I sit me down in a chair and explain to me what I’m trying to learn and what I hope will result. – Theresa Zook
  • Feel like everyone likes trying something new & shiny, and if you can forecast positive results responses are usually positive. – Noah Brooks
  • So important to set expectations, especially in terms of time the test needs to run. IE, not magic or immediate. – Heather Cooan
  • Explain that testing is the lifeblood of PPC and we are always trying things. Good to keep them in the loop. – Luke Alley
  • Show them the worst losses and tell them they are bad and should feel bad. Then bring up testing as the antidote…SOLD. – James Zolman
  • Back to LP testing being so beneficial, unfortunately it can be the hardest challenge to get resources to test them right. – Ginny Marvin
  • Are you the PPC Expert? Oh that’s right you hired me… So do what I say. Seriously though, trust is huge and educating clients is key to gaining trust. – Bryant Garvin

Q7: What mistakes/pitfalls would you advise newbies to avoid when testing? 

  • Statistical significance. Understand it or pay the consequences. – Melissa Mackey
  • Simple: Don’t set and forget. – Brian Gaspar
  • Don’t set it and forget it. Always set a date to review. Make sure results are significant. With data, no need for gut feelings. – Francis Shovlin
  • Hard to choose between “over-promising results” and “not establishing clear success/fail guidelines." – Theresa Zook
  • Do not decide too early. really easy to get a false signal and acting too early can tank performance. – Chris Kostecki
  • Don’t try to test too many changes at once. Small, incremental changes more often. – Jesse Semchuck (@jessesem)
  • Know the business and the audience or don’t even go there! – Heather Cooan
  • Soo many…1. Not tracking it correctly 2. Stopping too early/too late 3. Its OK if it fails and more. – Noah Brooks
  • Understanding that correlation does not imply causation. – David Kyle (@DavidKyle)
  • Failing to tell clients Google pressures us to test new features that don’t always “work." – David Szetela (@Szetela)
  • Test one thing at a time so you know what change produced the results. – Kasia Lorenc (@kasialorenc)
  • It’s really easy to make snap judgements. You’ve got to fail to win in ppc. – Jesse Semchuck
  • Sometimes you just win, no reason. Always make sure tracking is clean and consult the data! – Heather Cooan
  • Testing for the sake of testing just to show client something new. Not everything new is worth testing for every client. – Linda Delp (@lindadelp)
    • Just because it can be tested doesn’t mean it should. Prioritize. – Brian Gaspar
  • Client provides best insight, & has prob been in the industry a while, use them as a soundboard (bonus: it builds relationship). – Chris Kostecki
  • Don’t be afraid to try something different from everyone else, be willing to fail- admit & fix failures quickly. – Bryant Garvin
  • Get the scoop on what’s going on in other channels..synergy is your friend and the best place for testing fodder. – Heather Cooan


More PPCChats

Don’t forget to stay tuned for the next #PPCchat on Tuesday at 12 noon Eastern, 9 am Pacific and 5pm in the UK. Same Chat time, same Chat channel.


Check out the PPCChat Twitter list to see and connect with all current and prior participants.

• Matt Umbro (@Matt_Umbro)
• Paul Kragthorpe (@PaulKragthorpe)
• Luke Alley (@LukeAlley)
• Aaron Levy (@bigalittlea)
• Brett Stevens (@BrettStevens1)
• Brian Gaspar (@BGasper)
• Cassie Allinger (@CassieAllinger)
• Chris Haleus (@chrishaleua)
• Chris Kostecki (@chriskos)
• David Kyle (@DavidKyle)
• David Szetela (@Szetela)
• Elizabeth Marsten (@ebkendo)
• Eric Bryant (@GnosisArts)
• Eric Farmer (@click_eric)
• Francis Shovlin (@fmshovlin)
• Ginny Marvin (@GinnyMarvin)
• Heather Cooan (@HeatherCooan)
• James Zolman (@jameszol)
• Jesse Semchuck (@jessesem)
• Kasia Lorenc (@kasialorenc)
• Linda Delp (@lindadelp)
• Matt Dyer (@mattchstik)
• Melissa Mackey (@Mel66)
• Michelle Morgan (@michellemsem)
• Neil Sorenson (@iNeils)
• Nicole Mintiens (@Tregesy)
• Noah Brooks (@noahbrooks)
• Richard Fergie (@RichardFergie)
• Theresa Zook (@I_Marketer)

Streamcappin fromthe Cold Minnesota Tundra

This is a guest post by Paul Kragthorpe; WebRanking SEM Manager in Minneapolis, Minnesota, #PPCChat Streamcap Grabber, SEO Blog Author. Connect with me @PaulKragthorpe, and Google Plus.

Be Sociable, Share!

Tags: , ,

Leave a Reply

Your email address will not be published. Required fields are marked *

Recent Tweets

I am speaking at SMX East
PPCChat.co was rated one of the Best PPC Blogs by Boost CTR