So, Tuesday 21st January was the first Biddable World conference and it was just as successful as anticipated!
In this blog post, we (Holly and Tamsin) are going to give you our highlights and the tips that we will certainly be taking advantage of in 2014. We will also give you our Key Takeaway Message (KTM) from every talk! Like Jon Myers said – "it's all about the 3 letter acronyms, PPC, DSP… KTM"
Here are the highlights from the morning’s talks – come back tomorrow for the afternoon’s!
1. Jon Myers – Navigating the Biddable Landscape
Jon Myers took to the stage first and provided us with an excellent introduction to some present and future paid search trends:
- Clicks and impressions are on the decrease, CTR is on the increase, and cost is remaining constant.
- UK smart phone penetration is predicted to reach 75% in 2014.
- Take advantage of PLAs – their market share is increasing and they have a positive effect on CTR.
- Facebook Exchange ads are working well and allow us to target audiences at an extremely granular level. Facebook's strength is saturated in the platform of mobile and brings high CTRs and low CPCs.
- The world of display is becoming more auction-based and Real Time Bidding (RTB) is predicted to be at 50% by 2017.
- Audiences should be our new keywords! We can now determine our audience's journey, which allows us to generate personas and our keywords should reflect this! Build your data into these personas.
- Jon predicts that the world of advertising is also going to become more biddable and suggests that TV & Radio will join the auction and become biddable too!
KTM: The biddable world of media is growing, and at a very fast rate, which means more opportunities for us. Therefore, we need to identify these opportunities and exploit them sooner rather than later!
Heather Robinson dove deep into the Twitter advertising platform:
- She highlighted the three types of Twitter ads:
- Promoted tweets, which focus on increasing reach, awareness and engagement
- Promoted accounts, which increase followers
- Promoted trends, which focus on maximising exposure.
- Twitter's reach is currently very modest at 232 million, compared to Facebook's 1.15 billion. It is similarly modest in terms of a market share, with 1.85% compared with Facebook's 15.8%. However, its response is looking extremely promising. It has an engagement rate of 1-3%. (Facebook's CTR is 0.12%, but this is a different metric so cannot be directly compared.)
- If you identify the goals that you are trying to achieve and select the appropriate type of Twitter advert for you, target by device, geographical location, gender, user intent and interest and use the correct tactics, then Twitter advertising will work for you.
- Note that Twitter doesn’t actually know users’ genders – it guesses from their activity.
Some key tactics:
- Tweets that ask questions reap more replies.
- Images in tweets gain more engagement.
- Top tip tweets generate more favourites and retweets.
- Lead generation cards e.g offers, downloads and competitions are captivating.
KTM: Track your conversion in GA. Using the tweet activity report, identify the key performing tweets and implement them into a promoted campaign. Then, continue to run new techniques, report, improve and repeat. And remember, if you are reporting an engagement rate of less than 1% something is wrong! Address this and make changes to optimise!
Ben Harper highlighted his thoughts on Power Editor – Facebook's Bulk Editing tool – and told us how there are no excuses not to optimise properly using this platform….
- The Power Editor lets you copy, paste and amend very quickly so you can test variants of ads with different audiences and optimise.
- There are audiences accessible in the Power Editor and not the normal interface:
- Custom audiences, made of your own list of emails
- Lookalike audiences, letting you target people similar to your current followers
- In the USA you can use Partner Categories, which use offline purchasing data. For example you can target people who buy lots of children's cereal, or who have test driven a hybrid car.
- You also have more control over where your ad appears – your ad can appear just on Android devices, or just in mobile newsfeeds.
- You can also use Power Editor to get data – put in your competitor's name as an interest and vary the age range to see what their audience is like.
KTM: Use the Power Editor! Visit https://www.facebook.com/ads/manage/powereditor in Chrome
Nick Christian, a self-confessed PPC black hat had a really great point to make, and that was that PPC is different to other marketing channels because we aren't trying to manipulate people who have no interest in our product/service – the people we target in PPC are already showing an interest and are qualified to see our ads. This allows us to be super proactive and reactive!
- By 'black hat' he doesn't mean dodgy tactics like URL cloaking and arbitrage, but focusing on doing what actually works rather than 'best practise' – thinking less like a search geek and more like a maverick.
- He argued that there is no such thing as irrelevant traffic: traffic lies in a broad spectrum, where some traffic is more valuable than other traffic, but it's all of relevance! Everything has a conversion rate, even if it is really small and there's nothing you can't afford if the price is low enough.
- This led Nick to shift his focus towards keyword expansion, language expansion and device targeting expansion. He moved from a risk-adverse strategy towards a more risk-seeking strategy.
- This bold move saw Nick's CPCs drop, traffic increase and evenly split sales across all categories, all for the same cost.
KTM: Negative keywords shouldn't be neglected but adding new terms should be the focus. A risk-seeking strategy will be the one that gets you those exceptional results.
5. Richard Fergie – Is eBay Wrong About Paid Search?
Last March eBay released a report on their own test on their PPC's worth. Richard Fergie has been trying to recreate a similar test.
- There is a curve of profit versus spend – the two parts of PPC are changing the shape of the curve and positioning yourself on it.
- The 'right' position is going to be different for different businesses – but wherever you want to be, you need to know where you are now.
- You could get some data from Google's paid & organic report, but this shows data for searches where ads were not shown without saying why ads weren't shown – it may be biased to less commercial queries.
- One way to see the worth of PPC is the 'nuclear option' – turn it all off. But seasonality means you can't easily compare before and after.
- Instead go for the 'MIRV option' – nuke many small bits of your account. Then the rest of your account serves as a 'control' for comparison.
- You can't pause particular keywords – the remaining keywords may pick up the paused keywords' searches, unless you're very switched on with negatives.
- You could use Time Slicing – turn ads off for short periods. For example have ads running 1pm to 2pm, no ads 2pm to 3pm, ads 3pm to 4pm, etc. But this is tricky if you have a long buying cycle, unless you don't run the ads for a longer period.
- eBay used geographic slicing – they ran in some Nielson areas and switched the ads off in others.
- In the UK we don't have Nielson areas, so you have to come up with your own geographical groups:
- Ignore London – it's weird.
- Look at traffic at town/city level
- Normalise for population size
- Discard towns too small to see any change
- Make groups of towns/cities you'd expect to behave similarly. Control the risk to business by deciding how much traffic you're going to turn off – this depends on the size of the groups.
- Then the real test: turn off one of the groups and see how the drop in sales compares to the amount of sales usually attributed to PPC.
KTM: It's important to control risk, but it's also important to be correct when valuing PPC. This is also a way of seeing how good your attribution model is.
Tara gave us a deep valuation of the factors involved in quality score and I think an important point to keep in mind when thinking about your quality score is the fundamental difference between CTR and relevance. CTR is an indication of people voting with their mouse, whereas relevance is an indication of what Google considers relevant. Tara evaluated what quality score is and how it is optimised at different levels of the account:
- At keyword level, quality score is calculated on exact match search queries only and is based on historic CTR. You should use Broad Match Modifier (BMM) to boost initial impressions and add easy keywords followed by keywords that may not perform as well.
- If your keyword is new and doesn't have enough impressions, Google will use an industry average CTR for its quality score rather than its actual historic CTR.
- To get good ad level quality score, you should use extensions that are highly relevant, split test as soon as possible to identify ads with superior CTRs, use DKI in the headline and in the display URLs, and use variations of the keyword terms in the body of the ad.
- The account level quality score is an index of the overall historical performance of all of the keywords and ads in the account. So don't optimise some campaigns and ad groups and neglect others, as your account quality score is an overall measure. Also, optimise for location as much as possible as this contributes to your quality score and keep creating new campaigns, adding new keywords, and testing new ads with new, fresh content.
- Your landing page is a factor in quality score; it does not directly affect position but it does affect eligibility. It is all about relevance and user experience rather than CTR, and data such as bounce rates and time on site will assist in optimising this. What's pivotally important is matching the user's expectation and being transparent, so linking to a privacy policy and including contact information is favourable in the eyes of Google. It is important to remember that it is reviewed by humans as well as an algorithm, so it might be useful to ask an outsider to review the page and provide external feedback.
- At device level, the same factors are used to calculate quality score but a different calculation is used. Location is even more important for mobile relevancy so you can take advantage of that with mobile extensions that incorporate directions, a phone number and a contact page – all information that a mobile user is going to find more relevant.
- When measuring quality score you should multiply each keyword quality score by the number of impressions it has accrued, as this will provide you with a weighted impression share quality score.
- Use AdWords Scripts to keep track of quality score – they're quick and easy to implement. This one calculates at campaign, account and ad group levels, and this one is at keyword level.
KTM: Quality score builds its reputation from the first day a campaign runs live, so get your account structured well from the start. Be clever and set brand exact match keywords live first and let them build your quality score up for you, i.e. start by putting all of your eggs in the basket that you know is going perform really well. These terms can then support additional terms that may not immediately generate a high quality score.
Sarah illustrated the growing market share that PLAs are taking up in Google's paid search. She also touched upon Bing products ads and the potential that they are showing. She further provided us with some tips on understanding your shopper's habits and how to optimise PLA campaigns:
- Increase your budget.
- Use labels.
- Break out your feed by best sellers, margins or seasons.
- Make use of the accelerated delivery feature, if only for a four hour time period, to gain a real indication of performance trends and opportunities for growth.
KTM: PLA's share on Google search is increasing and pushing organic search further down the results page. So, take advantage of this movement and begin or optimise your PLA strategies now.
The post Highlights from the Biddable World Conference appeared first on White Noise.
Postări
Niciun comentariu:
Trimiteți un comentariu