The largest competitor and the fastest-growing competitor may deserve different places in your research queue. Start with three to five sites serving a similar buyer or task, then compare a common complete month. The example below turns those numbers into a decision about which site to inspect next, with a worksheet you can reuse.
Choose comparable businesses before comparing the numbers
A focused SaaS product and a broad software suite can solve one similar task while receiving traffic for very different reasons. The suite may attract visitors to documentation, a free utility or several unrelated products. Begin by writing the customer task you want to study, then inspect each candidate’s public offer. Keep a broad reference in the set only when you can explain what it helps you learn.
Hostname coverage can create another difference. One company may place its application on a subdomain while another keeps product use and marketing on the same host. Record the scope the estimator supplies and inspect the public URLs. If the scope cannot be aligned, explain the comparison as an observation of those websites rather than a complete comparison of the businesses behind them.
The most useful peer set often includes a close alternative, an established reference and an emerging product. Their roles help you interpret the numbers without forcing a single ranking. A larger site can teach distribution, a closer peer can teach positioning, and an emerging site can suggest a workflow worth trying. Replace any domain that adds volume but no relevant question.
Worked example: the largest site and the fastest grower
The following invented examples use the same two complete months and the same full-site estimate scope. Cedar grows from 100,000 to 130,000 visits. Maple grows from 10,000 to 18,000. Pine grows from 60,000 to 66,000. They illustrate how the question changes the ordering; they are not observations of named businesses.
Cedar adds 30,000 visits and grows 30%; Maple adds 8,000 and grows 80%; Pine adds 6,000 and grows 10%. Cedar leads in scale and additional activity. Maple leads in percentage growth. If the task is finding emerging workflows, Maple deserves inspection; if it is benchmarking distribution at scale, Cedar is the stronger starting point.
Compare channel volume as well as channel share
Suppose Cedar has a 20% organic-search share and Maple has a 70% share in the current month. If each share belongs to the same full-site estimate and period as its visit total, the implied volumes are about 26,000 and 12,600 visits. Maple depends more heavily on search, while Cedar has the larger estimated search-channel volume.
Write the multiplication as an approximation and retain its inputs. Keep a separate keyword-based search estimate in its own column. A modeled keyword total from another dataset can help you investigate search exposure, but substituting it for the channel numerator changes the calculation’s meaning.
Copy a working comparison sheet
Paste the tab-separated teaching example below into cell A1 of a spreadsheet. B and C contain the two matched months; D calculates added visits; E calculates growth. Format column E as a percentage. Replace the three synthetic rows with your own complete observations, and keep the two month labels, source and scope above the sheet.
Enter growth formulas only for rows with a known positive previous value. A missing baseline gets a blank growth cell and an “insufficient history” note. Sort D to study added activity or E to find proportional movers; keep column C visible in either view.
site previous_visits current_visits added_visits growth_rate
Cedar 100000 130000 =C2-B2 =D2/B2
Maple 10000 18000 =C3-B3 =D3/B3
Pine 60000 66000 =C4-B4 =D4/B4A shortlist share has a defined denominator
Cedar accounts for about 60.7% of the three sites’ combined current visits: 130,000 divided by 214,000. Describe this as its share of estimated visits within this shortlist. Your selection excludes other competitors and alternative ways to solve the task, so the calculation answers a bounded comparison question.
For this example, a usable conclusion is: “Study Cedar for distribution scale: it added 30,000 visits. Study Maple for an emerging workflow: it grew 80% from a 10,000 baseline. Next, inspect Maple’s search landing pages and check whether the increase persists.” Vaneform’s bulk export supplies a starting list; align its row dates before filling the worksheet. Stop collecting peers once you can name the next investigation for each shortlisted site.
What to do when the comparison refuses to line up
Suppose two domains have September estimates and the third only has August. For a current snapshot, keep the third row with its August label and leave it out of September totals. For a three-site comparison, look for August values for all three. If that common period is unavailable, preserve the partial comparison instead of carrying the old value forward under a new month.
If estimates are close, use the difference to choose what to inspect rather than declaring a precise winner. A modest gap can be less informative than a clear difference in customer fit or receiving pages. Open the products and the relevant search pages, then write down which observation would make you change the next action. That keeps the comparison connected to a decision.
For a repeat monthly review, preserve the original shortlist, the shared period and any source changes. Add a new competitor deliberately, because it changes the denominator of shortlist shares. Keep prior snapshots so that a change in the research set can be distinguished from a change in the sites themselves. The worksheet then becomes a record you can reason from, not just a fresh ranking each month.
A report’s update timestamp and its measurement month may also differ. A record fetched in October can still describe August visits. Put the measurement month in the comparison columns and the retrieval date in a separate note. When a refreshed report changes an older value, preserve which version supported the previous decision. This makes a later review understandable without assuming that every newly retrieved number belongs to a newly completed month.