Clicky

X

Subscribe to our newsletter

Sign up for Newsletter State of Digital
* = required field
Daily Updates

Above The Line, PPC and SEO – It’s a “Brand” New Game

12 July 2011 BY

First up, an apology. This is less of a informative post and more of an open-ended question to all marketers, and not necessarily for just those that focus on online or search marketing. I’m looking at brand, budget, media and ROI. And testing. Lots of testing.

Being an SEO at OMD, a full service media agency born from a purely offline media buying and planning service, for the past 11 months has been an eye-opening experience. Only half the people in the agency work in “digital”, so getting down on a campaign level as far as  “search” or even more granualar “SEO” can be tricky.  I hold my hands up, I understand next to nothing about how offline or “above the line” marketing works. It’s too easy to become siloed and work almost completely separately. You get so used to assuming that my link building has no relation to their outdoor advertising that integration is something that needs hard hard work, daily hard work, in order to make it happen.

Coming from a background of small specialised digital agencies, learning to take a step back and look at the bigger picture has been a slow process. Understanding the world of big brands, budgets and wider visibility is taking longer still. However I think the future of SEO, search marketing or digital marketing in general will require this. Having said that, I’m also learning that sometimes the only way to get to the bigger picture is to start with the small things and learn to look outwards.

As a full service agency, we often control budget across a variety of specialisations which we need to internally break down according to how we believe it is best spent. From digital advertising to affiliates to newspaper advertising to TV – how is it possible to gauge which creates the best ROI across a brand?

As an SEO, this is a twist on the age-old PPC vs SEO argument which has taken some huge leaps. It is much more than simply search marketing that we should be considering. But let’s start there…

Recent examples of some core questions being raised on a client account:

- Should we be bidding on pure brand terms using PPC?

- Should we be bidding on any brand terms (extended, misspellings etc.) using PPC?

- What about when competitors are bidding on our brand?

- Will SEO fully compensate for the extra visibility and manipulable messages PPC offers?

- How do we test this considering all the seasonality variables involved?

- With the amount of money spent on brand terms with PPC, would it be better spent being pushed back into ‘above the line’ techniques?

- How do we measure this?

- Should we rely only on SEO when we are pro-actively pushing more ‘above the line’ marketing offline or should we use PPC to get across more up-to-date messages if a person searches on our brand term?

The list of questions goes on. I would love to say I have all the answers to these. Sadly this is not (yet) the case. How are we going about trying to get these answers? Thank the lord for Google Analytics!

Types of Tests Being Undertaken:

- Switch off pure brand bidding completely, monitor organic traffic.

- Switch off varieties of brand bidding, monitor.

- Use search volumes and GA to estimate total clickshare.

- Use tools such as Hitwise or AdGooroo to estimate traffic to competitors.

Methods of testing:

- Geographically target similar locations with brand on while targeting others with brand off and directly compare on same dates.

- Brand on for one month followed by brand off the next month to enable more accurate keyword search volume/clickshare estimations

- Brand on for one day followed by brand off the next, alternating for six weeks then calculating the overall traffic difference to avoid seasonality issues

- Brand on for one week with heavy offline advertising then repeated 6 months later during periods of similar traffic/seasonality.

KPIs to measure:

- Increase in SEO traffic when PPC brand bidding is switched off

- Incremental increase in visits from PPC when brand is switched on

- ROI of these incremental visits

- Estimated clickshare of total search volume of brand keyword

- Estimated clickshare of competitors on total search volume of your brand keyword

- Direct vs SEO visits when offline marketing is heavily pushed

- All site traffic, direct, SEO and PPC when offline marketing is heavily pushed with brand bidding switched on vs brand off.

Results?

As you can guess, results are as yet inconclusive across the board. As you can see from the methods of testing, this is going to take months to get to a place we feel confident in. I’m looking forward to sharing these with you when I can. However I’d be interested to hear others’ opinions on this sort of testing, whether it has any value and if so the kind of tests others have undertaken.

One piece of advice I would share is to make these tests as scientific as possible. Formulate hypotheses, methods, agreed controls and KPIs before starting ANYTHING. And check, double check and triple check your Google Analytics is 100% correctly set up before even thinking about the next step. Trust me, we learned the hard way. Invalid tests after weeks of waiting are not a lot of fun.

Look forward to hearing your thoughts.

Thanks to riverpublishinggroup.com and http://blog.ineedhits.com for the photos.

AUTHORED BY:
h

Originally from the UK via France and Malaysia, Annabel Hodges is a digital marketer with long experience in the industry now residing in Sydney. She heads up the Digital Marketing at Next Commerce, working across an array of products, channels and brands.
  • http://richardfergie.com Richard Fergie

    I’ve blogged on the PPC side of this at SEOptimise http://www.seoptimise.com/blog/2010/03/is-ppc-canabalising-your-brand-traffic.html

    Two other ideas that I’ve had since then:

    Use AdWords Campaign Experiments to segment people in a more random way. Reducing the bids to 0.05 in X% of auctions and seeing the effect on sales is the closest thing I can come up with to a holdout test.

    Split test remarketing tags. Show X% one tag and Y% another. Advertise to one but not the other. See what %age would have converted anyway and how much incremental value retargeting is driving.

    I’m interested in hearing how you account for seasonality and the effect of ongoing work when reviewing these tests. The problem I’ve always had with long running tests is that I end up having to do some work in the test area which ruins everything.

    • http://www.annabelhodges.com/ Annabel Hodges

      Awesome, thanks Richard. Useful post to read. I definitely like the idea of Campaign Experiments – worth a shot. I think the main issue here as you mention is time and having the luxury of leaving a test to run for a significant period of time overcoming seasonality. We are of course always under pressure to reduce spend on brand ASAP as the hypothesis will always be that it is poor value. What we are trying to do is run these tests at multiple times of the year when we know traffic is fairly stable, and compare.

      Incremental traffic from PPC on brand terms I think is particularly crucial when considering other marketing activity being undertaken at the same time. e.g. consider a sale/promotion advertised in the paper. If a person searches on that brand and sees only the standard organic snippet making no mention of the sale, is that less likely to convert than also including an up-to-date message within a PPC ad?

      Remarketing tags is a nice idea although I think a little more sidelined from what I’m currently looking at, i.e. direct impact on reduction of core PPC budget in order to move savings over to other marketing streams.

  • http://www.turn-key.co Stuart P Turner

    Hello Annabel,

    Really interesting post, I’ve just made a similar move from a purely digital focused agency to work in an integrated marketing agency (it’s not quite Revenge of the Nerds but you get the idea).

    Anyway… you mention that “We are of course always under pressure to reduce spend on brand ASAP as the hypothesis will always be that it is poor value.” which piqued my interest. I was having a conversation yesterday morning with the Head of PR here about this subject, in the context of clients vs. professionals writing website copy.

    I know it’s a tough climate at the moment and everyone is watching the pennies, but (imo) if anything I think that requires more spend on brand. I say this because it’s the established, trusted brands that can weather these kinds of economic storm more comfortably than the small guys.

    The reason for that is they already have brand equity, so while you may see a lower return on a brand campaign if you look at it in isolation, what I would expect to see is a holistic uplift within a multichannel campaign, when brand features in all channels.

    I’ve never managed to successfully test this as I haven’t had the luxury of access to full marketing performance data (before here) however Adgooroo/Hitwise are great from online analysis. If you can steel yourself for the challenge Atlas is (supposed) to be awesome once it’s set up.

    The most solid testing I managed to do was the impact of TV on brand conversion – however! This test was skewed hugely by the strength of the offers featured in the TV, and changes in channels/times on which the ads were shown. As a result the data was inconclusive, but I’ve got some ideas about how to expand this test if I can find a client willing to get involved…

    • http://www.annabelhodges.com/ Annabel Hodges

      Mr Turner hello, and thanks for your comment. In this case, I think I’ve perhaps done the client an injustice with my lack of clarity on brand spend. They are keen and willing to spend on brand recognition, visibility and reinforcement in general however they (I think correctly) question whether putting a large amount of budget in brand bidding on PPC is the best way to do this.
      e.g. If you spend a significant chunk of money on PPC solely on your brand, would you potentially be better off re-investing that in a targeted newspaper campaign to attract an audience not already looking for you? Even if the amount of you get for your money is less, could it prove a better brand boost in the long-term?

      Totally agree with the problem of tests being skewed by offers. This is one of the biggest issues we face. In general, budget is only being fed into above the line streams when there is an offer on. Therefore how can the impact of this be measured against other non-promotion times of year?

      The questions continue….

      • http://www.turn-key.co Stuart P Turner

        My pleasure Miss Hodges.

        I see what you’re saying, a valid question to be fair.

        I suppose I’ll have to answer your question with more questions:

        Are there competitors bidding on your client’s brand terms as well?
        While this is a notoriously low conversion approach to PPC would your client lose brand visibility and see an overall drop in traffic / conversion / engagement as a result? The variables to consider and tracking you would need to run a test like that would be immense…

        I suppose you could run a direct comparison of brand + offer vs. brand + no offer however, to ensure it was as accurate as possible you would need to ideally ensure:
        - you are not in a seasonal peak / trough; the fact that you can’t run them concurrently is already going to affect accuracy
        - there was enough of a gap for the offer buzz to die down
        - ad copy and placements remained as identical as possible (minus the offer obv.)
        - brand perception didn’t change over the period of the two tests (e.g. you don’t want a NOTW or BP style disaster)
        and most crucially
        - all other marketing either remains the same or, if other channels feature the offer they follow the test parameters as well

        That already sounds like a nightmare to set up and run.

        That basically doesn’t help at all does it?

  • http://www.esterling.co.uk/ eSterling

    Hi Annabel, thanks for the really interesting post. It’s very insightful to see other SEO professionals asking questions and looking for answers from others. When weighing up your options, trial and error is sometimes the best way, because there’s no one-size-fits-all option. I would be interested to know how you get on in a few months time… make sure you let us know.