Posted by & filed under PPCChat.

Guest host Sonika of Seer Interactive joined host Julie F Bacchini for this week’s interactive PPCChat session where experts talked about the process they use to analyze the data, what tools and automation are they leveraging to analyze large datasets, and more.

Q1: What’s your go-to process for analyzing query data? For example, I combine x data sources and [fill in the blank]…

At @SeerInteractive, we start by combining paid search query data with organic keyword ranking data from our partner, @AuthorityLabs to find gaps in performance. Think: good organic rank + low paid CVR or bad organic rank + expensive CPCs. @sonika_chandra

I feel like I spend a lot of time lately comparing queries across campaigns or even ad groups to see if they are triggering in more than one place. Mainly use Excel for this after pulling the query data from G Ads. @NeptuneMoon

This allows us to find search themes that we can either expand upon, or optimize if they are low performing. @sonika_chandra

I always start with the search query analysis that @Adalysis serves up. Best place to go for low-hanging fruit (good and bad). Then I pull raw from G Ads into Excel and run some pivots. @beyondthepaid

I pull in the matched keyword to see if there are any accidental duplicates + assess if I’m investing in the right keyword champion. I’ll also segment by conversion action to get a sense if the query is driving top or bottom of funnel conversion actions. @navahf

I’ll also check search term data against organic search console data to see where there’s overlap and if it’s possible to target high value terms that the organic team is struggling to rank for/overcome the reason for their bounced traffic. @navahf

We use the search query data we can get from Google Ads and from Google Analytics, chat with our SEO team and see what information they can add by collaborating with our Media team! We’ll also use n-gram scripts to find additional terms to target or remove. @snaptechmktg

I’ve been pulling query data from ad platforms (campaign + ad group level to find overlap) + GAnalytics to understand what’s working/what’s not. When I can, I like to match up paid queries with organic data as well to find gaps/areas where they can support each other. @jennifer_lash

I do also like to review Search Console data periodically to see what I might be missing on the paid side of things. Connect Google Search Console to your Google Ads accounts if you can! @NeptuneMoon

One other note: I will always check my historical data bias against Google trends to see if a given keyword represents how most folks are searching or if the wind’s blowing another way (especially useful now that we don’t have full search term data. @navahf

We look for gaps using Google Search Console data and find words that we rank for organic but not paid, paid but not organic We also use the keyword valuetrack parameter and some Google Ads scripts to find anomalies. @jimbanks

At Seer, we also use these findings to help in developing Paid Social ad copy and landing page strategy. @nicoleinabox

Q2: What are your primary success metrics for your queries? CVR (conversion rate)? CTR (click-through rate)?

It’s my favourite answer in marketing (I even have the shirt). “It depends!” Generally, conversion rate data combined with volume metrics. In lower volume accounts we will also look at higher funnel softer conversions and in e-comm ROAS. @amaliaefowler

Is term relevant? Does it has potential? Low CVR: OK w/Low CPC CTR: change w/AdCopy @GoogleAds goal: control searches your ad show up for and pay what it’s worth to you Find relevant Term not in acct, add as Exact, write ad, send to best page & bid what it’s worth. @yaelconsulting

Our primary success metrics from queries vary by account type and goal! In general conversion rate and ROAS are key – but we also use the Top Conversion Paths report in Analytics to ensure we aren’t missing any steps. As @amaliaefowler says often “It depends”! @snaptechmktg

I look at the conversion rate/cpa for the conversion action the query drove. A super low CPA with a high conversion rate for a branded query is cool and all, but kind of expected. Becomes much more interesting when that happens for a breakout search term. @navahf

Classic answer… It depends. If it is a conversion related keyword then it will be CPA, and we also look at search volume and abs search volume. But if the objective is brand awareness we might be more interested in CTR. @jimbanks

When scanning for irrelevant query themes or mismatched intent, we start by breaking our query list into n-grams. (Easier to scan!) We drop our data into @MSPowerBI and analyze the average CPA of the n-gram or bi-gram as compared to the account average CPA to ID outliers. @sonika_chandra

if we have tracking attribution installed we most definitely look at CVRs if no attribution then CTRs. @PPCKenChang

*it depends* usually both CVR + CTR along with ROAS/CPA where applicable, with each being weighted differently based on the account: volume/types of conversions that are coming in – it also differs for lead gen v. ecomm. @jennifer_lash

We are also looking for outlier terms that might be unicorn terms and modifier terms that could be useful as negatives. So occasionally the data we seek is not governed by a specific metric but has an underlying practical purpose. @jimbanks

I tend to work backwards from the client’s desired outcome + the intent level of the audience. Conversion Action Sets are wonderful for this (and really, really under-utilized). @DigitalSamIAm

The other big one for me is net profit. Intermediate (and more accessible/widely used) metrics like ROAS are like gasoline – potentially useful + potentially dangerous, just depends on who has their hands on it. @DigitalSamIAm

For me, the core is “Should we be showing for this search term?” Followed by “Should this ad/landing page show for this term?” I think the hard metrics tell us something different than common sense in a lot of cases. @ferkungamaboobo

Other stuff (CTR, CVR, QS, whatever) tends to be more like an RPM gauge to me vs. a KPI — it is useful for understanding where there might be opportunities, but rarely is solving for it the right life choice. @DigitalSamIAm

Q3: At Seer, we don’t live in silos, are you leveraging other platforms’ search term reports? Organic data? If so, how?

In addition to regularly using organic ranking data – recently, I have been playing with search term reports out of @AmazonAds to join with our Google Search term reports with the goal of finding trends between the 2 platforms. @sonika_chandra

At Snaptech, we’re big believers that even if we’re operating on one platform for our clients that we need to be leveraging the data from a holistic perspective and working with any third-party providers. We’ll work with the social team around messaging (1/ @snaptechmktg

The search terms report from Microsoft Ads has been more insightful lately. @PPCKenChang

YES!!!!!! The easiest way to break down silos is to build in audience cultivation workflows (analytics/utm parameters), as well as message mapping according to traffic source. When all channels work together, it’s much cheaper (time/$$$) to get winning results! @navahf

Don’t forget that most platforms that allow you to run ads will provide a query report that you can leverage to understand your broader market! Think outside the Google/Bing box. @sonika_chandra

I do love me some Google Analytics, and as I mentioned in A1, Google Search Console (which you can also connect to Analytics). I use Keywords Everywhere and do some manual searches from time to time to see what else is suggested/related. @NeptuneMoon

Another useful function is to build in automatic data sharing (search term data and search console data) between teams. If the organic content team knows something converts, they can be empowered to invest. Same with high value but low search volume terms. @navahf

We love the valuetrack parameters we get with Microsoft Ads. We add : {Keyword} {QueryString} {MatchType} {BidMatchType} It fills in a lot of gaps in search data for us in GA and @luckyorange @jimbanks

I like looking across all ad platforms that the client is on, in Google Analytics at both paid + organic query data, and with my SEO counterparts to understand what data they have…also the SERPs for research + “other people searched for…” suggestions. @jennifer_lash

We also use the audience valuetrack parameter because the context of the term + the audience is often what gets end users over the line and converting. So being able to adjust a list modifier up or down depending on the value it adds has been impactful @jimbanks

Q4: What tools and automations are you leveraging to analyze large datasets at scale? On average, how long does each analysis take?

Like I said in A1, @Adalysis does a great job surfacing key SQ data. Scripts do too. Using some type of tool or automation is crucial for accounts with thousands of keywords! @beyondthepaid

Are you all sick of hearing about @MSPowerBI yet? It is my most frequently used tool to analyze large datasets at scale, but we have recently been using @Optmyzr & @Kenshoo to help us identify trends in our data as well! @sonika_chandra

Bigml and Google collab @PPCKenChang

Special shoutout to @andchristina & @wilreynolds for getting me started in PowerBI. A total gamechanger for analyzing paid media data. No more sifting through UI filters & dealing with clunky interfaces for me. @sonika_chandra

Data studio and google sheets (excel) are great free options. In terms of time: it depends. When I’m doing client work, analysis is maybe 1-2 hours per week per client to make sure they’re on the right path. When I do #ppc content, it’s at least 10-15 hours. @navahf

We are trying to do more data warehousing using BigQuery, Google Cloud, DataStudio With ATT and Privacy Sandbox it is going to fall on the advertiser to manage and understand the interactions of users on their first party data, we can’t rely on third party any more. @jimbanks

Optmyzer is a good one I have a nerdy obsession with Power BI for larger data sets, it’s been really great for me for looking at a lot data from different perspectives at the same time without a whole lot of time consuming legwork. @jennifer_lash

So what’s large? For most clients/accounts, something like @Adalysis is a great solution. For larger datasets or to join datasets across accounts, I like R (handles the data science piece) + Tableau (to visualize it for the client). @DigitalSamIAm

Optmyzer, Google Data Studio, Excel/Google Sheets, and I have a nerdy obsession with Power BI for larger data sets, it’s been really great for me for looking at a lot data from different perspectives at the same time without a whole lot of time consuming legwork. @jennifer_lash

It’s important to factor in what we’re analyzing. Keyword/search term analysis doesn’t take nearly as much time as ad creative analysis or buyer persona creation/benchmarking. Shoutout to all “ops” marketers who put up with all our wild and crazy requests. @navahf

Q5: When analyzing your query data, what insights and opportunities do you typically find?

For one thing, I see how crappy some of the matches are. It’s shocking, really. @beyondthepaid

So many! Questions that can be great blog or FAQ content! Terms that we can negative! Terms that we can turn into their own campaigns and target individually! @snaptechmktg

Lots of them – insights into how Google reads our site (and how terrible it is at matching intent). We also find content ideas, terms we can negative, terms we can breakout into their own campaign. I’m a hard no with broad match but sometimes its eye opening. @amaliaefowler

They usually fall into four buckets: (1) gaps + misalignment in our current strategy; (2) areas where we’re spending w/ suboptimal return; (3) user behavior + preference evolutions; (4) marketplace changes. @DigitalSamIAm

I find what I’m looking for (the mindset you analyze your data with absolutely influences the outcomes): 1. Am I looking for spikes in cost (& whether they’re warranted because they lead to better customer value)? 2. Am I looking for inefficiencies/winners? @navahf

DISCLAIMER: most “insights” people find are noise, not signal. So the first step is to ensure you’re looking at a robust + representative (or close to it) sample size. Just because Term X has a 30% conversion rate doesn’t mean it’s the new hotness. @DigitalSamIAm

Lately, NEGATIVES to add. But also you can learn about intent and substages of a searcher’s journey with the queries. Find data points and items you should be sure to address in your content. @NeptuneMoon

Typically it’s one of three camps. Positive (keyword is a winner/keeper) Negative (keyword should be excluded at campaign/account level) Content (keyword might be good for content to support a pillar page) @jimbanks

Once I know what I’m looking for (and I’ll try to limit data adventures to a single objective to maintain quality of outcome), I can apply realistic benchmarks to the search term data. @navahf

Negatives and a lot of relevance insights lately + how Google’s reading the landing pages for content ideas, account expansion opportunities, deeper insight into searcher intent. @jennifer_lash

This is my “it depends” answer. Some quick wins: 1. Root terms that we can add to an account wide negative list. 2. High performing themes to expand upon. 3. Root terms that are crushing it on PPC that we can write organic content around. 4. Opps to improve ad copy. @sonika_chandra

If I see queries that perform well on shopping but don’t exist on the search side, then I’ll normally test them out. Also find opportunities like negative keywords that need to be added. Lastly I look at what does well with paid but we’re not ranking for organically @BrettBodofsky

Q6: It’s been 6+ months since the Google Ads search term report update, are your accounts continuing to be impacted? What creative strategies/tactics have you leveraged to circumvent the query data loss?

Our accounts are definitely impacted, mostly being low-volume we see even conversion queries not reported on. Our media team is leveraging Google Analytics reports more, and we’ve really tightened up all our match-type targeting. @snaptechmktg

This ties in with more of a leaning in to not trying to control things that are outside of my scope of control. I find the machine learning doesn’t do a bad job, it’s freed up some time, so we can focus on building out other channels and sub-channels better @jimbanks

We still see a lot of data loss, even from converting queries. We’ve adapted all our processes to use what tools we do have (like Analytics) and leverage negatives more. @amaliaefowler

Viewing them in Google Analytics, because they are all there. @DavidKyle

Definitely being impacted, some very significantly. For now, I can piece together at least some of the missing query data in Analytics. Not ideal, but better than nothing? Interested in what others are doing, for sure! @NeptuneMoon

Yes, I use a script that identifies the % of queries missing from Google’s reports, some clients are seeing 60%+ missing terms…. I’ve been relying more on Google Analytics query reports. Very interested to hear what others are doing! @jennifer_lash

Without getting too deep into the update & it’s implications – we have been working hard to find alternative data sources to help us understand search trends. This includes platforms outside of Google, as well as using Audience & Placement Data to find “bad intent.” @sonika_chandra

UTMs! Even if I can’t see the full query, I can make intelligent guesses about what the query might have been based on the match type: ({lpurl}?matchtype={matchtype}&device={device}&keyword={keyword}) I wrote a @sejournal post on this: @navahf

I can’t wait to hear everyone’s answers to this one! @sonika_chandra

Accounts are surely being impacted by the obscuring of search term data. Nothing is worst than not being able to see a search term that lead to a conversion. Excited to hear how others are dealing with this. @BrettBodofsky

Make friends with other departments. Get insights into why a lead dies/items are returned. Create content that speaks to human concerns and the semantics of the query won’t matter (although we might all find ourselves using phrase match now). @navahf

Q7: Do you have any questions for @sonika_chandra and @SeerInteractive about query data?

What’s your best resources for learning Power BI? (also sorry if you answered it already somewhere) @amaliaefowler

Are you doing anything now to prepare for either keywordless search targeting or at least the continued demotion of keywords as a targeting factor in search ads? @NeptuneMoon

What’s your advice to smaller folks who don’t have the data sets that you have at your disposal? How can #ppc folks create @SeerInteractive level data/content if they have 1-10 brands? @navahf

PPCChat Participants

Related Links

Leave a Reply

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

Website Protected by Spam Master