Doing it yourself vs hiring a provider
We sell the second one. Here is the honest version anyway.
We are not a neutral party on this question, so the useful thing we can offer is specificity: which parts of the work a team can genuinely run itself, and which parts fail in-house for structural reasons rather than for lack of skill.
The short version: measurement and site fixes are very doable in-house. Source placement is where in-house efforts usually stall, because it needs relationships and a publishing cadence that one person cannot maintain alongside another job.
Where they actually differ.
| In-house | Provider | |
|---|---|---|
| Baseline measurement | Very doable — an afternoon, no tools required | Faster and more consistent, but not a capability gap |
| Site readability fixes | Doable if you have an engineer with time | Usually faster because the spec is already known |
| Content production | Doable, and often better — you know the subject | Useful mainly for volume and consistency |
| Source placement | The usual failure point — needs relationships and cadence | The main thing being bought |
| Sustained re-measurement | Starts well, lapses after two or three rounds | Contractual obligation, so it keeps happening |
| Cost shape | Staff time, hard to see on a budget line | A visible line item |
When to pick In-house
- You have someone who can own it for a few hours a week and will still be doing so in six months.
- Your category has thin competition — the candidate pool is shallow enough that good content alone can win.
- You want to understand the mechanics before buying anything. This is a good reason and we recommend it.
When to pick Provider
- Your baseline shows answers leaning heavily on third-party sources you have no route into.
- Nobody internally can hold the re-measurement cadence. This is the most common real reason, and it is a legitimate one.
- You need the work to be auditable — raw records, publishing logs, a method someone else can check.
Questions teams ask before they start
What is the minimum we should do ourselves regardless?
The baseline. Write down 15 to 30 real customer questions, run them across the platforms, record who gets named and what the answers cite. It costs an afternoon, it is the only thing that makes any later claim checkable, and it is your main negotiating leverage with any vendor.
What does it cost to run it in-house?
Staff time rather than an invoice: roughly half a day per measurement round, plus engineering time for the site fixes, plus whatever the placement work takes. The last item is the one that does not fit into spare hours.
Can we do the measurement ourselves and buy the rest?
Yes, and it is a reasonable split. It also gives you an independent check on any provider's reporting, since you can re-run the same frozen set and compare against their raw answers.
What should we ask a provider before signing?
Three things: how results are measured, whether the question set is fixed for the term, and whether raw answers are provided each round. A provider unable to answer all three cannot be held to a result.
How do we hand over if we start in-house and then outsource?
Hand over the frozen question set, the raw answers from every round already run, and the dates. That history is the reason a later provider can be held to a comparison rather than starting a fresh baseline.
Is there work we would not recommend outsourcing?
Substantive content in regulated categories, and the final confirmation of any published claim. Both carry consequences that land on you rather than on a provider, so both should stay under your control regardless of who writes the draft.
Other comparisons
Find out where your brand actually stands in AI answers.
An AI visibility diagnosis across the platforms your customers use, on your real questions. No charge for the first look.