Competitor ppc analysis

Start with the competitive question, not the data dump

Analyst-led competitor PPC analysis starts with a specific question: 'why did our impression share drop on keyword X', 'who's the new bidder in this auction', 'why are our CPCs trending up'. Without a question, you produce a data dump nobody reads. Write the question at the top of your analysis doc; every subsequent step references it. This is the single discipline that separates useful competitor PPC analysis from busy work.

Pull Auction Insights filtered to the question

Google Ads → Campaigns → Auction Insights. Filter to the keywords in your question; filter to the time range the question covers. Export to CSV. The columns to read: impression share (yours and competitors'), overlap rate, position above rate, outranking share. These four metrics together answer most competitor PPC questions before any external tool is touched.

# Reading the 4 key Auction Insights metrics:
# Impression Share: % of auctions where the advertiser appeared
# Overlap Rate: % of your auctions where this competitor also appeared
# Position Above Rate: % of overlapping auctions where they outranked you
# Outranking Share: combined measure of who 'won' the head-to-head

Identify the specific competitor(s) the question is about

From the Auction Insights export, identify which competitors' metrics changed in ways that explain the question. If your impression share dropped 15 points, look for competitors whose impression share grew by similar amounts in the same window. The math doesn't always close cleanly (auction dynamics are non-zero-sum), but the shifts point at the competitors driving the change.

Pull each suspect competitor's ad copy and bidding history

In SpyFu or SEMrush, pull the suspect competitor's ad copy and bidding history over the question's time window. Look for: new ad-copy variations introduced in the window, keyword expansion (new keywords they started bidding on), seasonal patterns. Often the competitive shift correlates with a specific change the competitor made - a new ad campaign, a price test, a category expansion.

Validate the hypothesis against multiple data points

One correlation is a hypothesis, not an answer. Check: (1) did the competitor's landing pages change too? (2) did their estimated CPC shift? (3) did other competitors' metrics change in the same window or just this one? Multiple aligned data points harden the hypothesis. A single correlation can be coincidence; three correlations are signal.

Translate the analysis into a specific recommendation

The output of competitor PPC analysis is a recommendation, not an analysis. Format: 'Question. Answer. Recommendation. Risk if we don't act'. Two paragraphs maximum. The competitor PPC work that earns its weight produces clear recommendations that bidding owners can act on inside the same week - 'increase bid by 12% on keyword set Y' or 'introduce price-led ad variant matching competitor X's pattern' or 'accept the lost share and reallocate budget to keyword set Z'.

Archive the analysis with the question for future reference

Save each investigation - question, data, recommendation, outcome - in a shared analysis log. Six months from now, when a similar question arises, the log shortcuts the investigation. Over a year, the log becomes the team's competitive intelligence memory. Analyst-led competitor PPC analysis compounds through the archive; one-off analyses don't.