When Your Multi-Venue Surveillance Has a Blind Spot in the Handoff Between Sites

You've got cameras at five sites. Maybe ten. They're all recording, but when a person or vehicle moves from one property to another, the trail goes cold. That gap—the handoff between venues—is where incidents slip through. And it's surprisingly common.

A 2023 survey by the Security Industry Association found that 42% of multi-site organizations reported at least one security event where footage from different venues couldn't be correlated in time. The cause? Systems that don't talk to each other across the handoff. This article is for the person who has to fix that: the security director, the IT manager, the facility lead who's been told to 'make it seamless.' We'll walk through the options, the trade-offs, and what actually works on the ground.

Who Needs to Decide—and by When?

Identifying the decision maker: security director vs. IT vs. operations

Most teams skip this step—and that’s where the problem starts. The security director usually owns the surveillance outcome, but IT controls the network, and operations manages the physical sites. I have seen this triangle stall a handoff fix for six months while each person waited for the other to blink. The real question: who holds the budget for the integration? That person must own this decision, not merely consult on it. If your org chart splits video systems under security and network under IT, the handoff gap lives exactly in that white space nobody filled. Worth flagging—the operations lead often knows which handoffs are dropping video, but they rarely have authority to approve a cloud connector or a new edge appliance. So before you talk architecture, talk responsibility. One concrete fix from a project I worked on: we forced a single sign-off from the person who could write the PO, then looped IT in as technical advisors. That cut decision time from four months to two weeks.

Timeline pressure: before lease renewal? after incident?

The calendar drives this harder than any technical requirement. If you're mid-construction on a new venue—or worse, expanding an existing site—the handoff gap must be closed before the next camera goes live. Why? Because retrofitting a handoff after cable runs are buried and switches are locked in rack cages costs 2–3x more, and that assumes you can get a maintenance window. Not yet convinced? Think about lease renewals. Most multi-venue operations sign 3–5 year leases, and landlords restrict structural changes after move-in. I have seen a security director lose their window to run a dedicated fiber link between two buildings simply because the lease renewal paperwork was already stamped. The catch is that incidents—a break-in at the boundary between two sites, a theft that passed through an unmonitored handoff zone—create urgency but also invite blame. Better to decide before the post-mortem report lands on someone’s desk. A practical benchmark: if your next site integration is less than 90 days out, start the handoff conversation this week. That sounds tight because it's.

Consequences of delay: coverage gaps during construction or expansion

What usually breaks first is the seam between two sites that are supposed to share a live feed. You install cameras at venue A and venue B, but the video handoff between them uses a patchwork of VPN tunnels and employee-initiated transfers. Fragile. Wrong order there—you're building a chain where each link expects the other to call first. Meanwhile, the construction crew is moving materials across that seam, or the expansion adds a parking lot that sits exactly in the handoff blind spot. Delays compound fast: one week of waiting on a budget sign-off creates three weeks of technician scheduling lag, and by then the site is framed and drywalled. I fixed this exact problem for a retail chain that added a new building 200 feet from their existing warehouse—they waited until the roof was on to ask about handoff latency. The result? A 14-second gap between when a person left one building and appeared on the next camera system. That hurts. A

'Fourteen seconds is the difference between catching a face and catching a license plate. We caught neither.'

— Security Manager, regional logistics firm, after a loading-dock theft during expansion

Don’t let your handoff strategy become a footnote in the construction punch list. Decide early, assign the budget, and accept that the decision will be imperfect—but made before the concrete dries.

Three Ways to Connect Sites: Cloud, Edge, or Hybrid

Unified cloud platform: all video to one cloud

This is the simplest story to tell—every camera at every venue streams directly to a single cloud service. Verkada, Arlo for Business, and Eagle Eye Networks sell this. You get one timeline, one search bar, and the handoff between sites is handled server-side. The metadata (face detections, license plates, motion events) arrives pre-merged. Two venues? Same pane. Ten? Same pane. That sounds like the dream until you look at the bandwidth bill—each camera pushing 24/7 H.264 eats through a 1 Gbps circuit like paper through a shredder. Worth flagging: if your internet backbone drops for twenty minutes during a handoff window, you own a dead period. Not a gap. A hole. The catch? Latency. That cross-site search you want—it takes ninety seconds to scrub two weeks across four venues. Acceptable? Depends on whether you’re chasing a yesterday theft or a live tail.

Edge recording with central aggregation

Local DVR or NVR at each site. Hikvision, Axis, Bosch—these giants prefer this. The cameras write to local hard drives. A VPN tunnel stitches the sites together, and a central VMS (video management software) pulls metadata summaries hourly. The handoff logic here is brutal but honest: each site’s timeline is independent. No shared clock drift correction—you fix NTP sync yourself. I have seen teams build their own metadata bridge using REST APIs, only to discover site A’s timestamps are six seconds ahead of site B’s. That six seconds killed an investigation. The upside is resilience—internet drops, you still record. The downside is your operator now has to mentally splice two timelines together. You can buy aggregation tools from Milestone or Genetec, but they cost as much as a midsize sedan. Most teams skip this: they forget to align the retention policies. Site A purges after 30 days, site B after 90. Handoff query fails for anything older than 30. That hurts.

Hybrid: local storage for latency-sensitive cameras, cloud for cross-site search

This is the messy middle that works. Keep your critical cameras—entrances, cash rooms, loading docks—recording locally. All other streams go cloud. The handoff metadata lives in a cloud index: event logs, object classification, and clipped thumbnails. When an operator searches, the cloud returns a map of “go check device 7 at site B, timestamp 14:22:03.” The actual video stays local. Why does this matter? Because the handoff gap shrinks to the metadata sync interval—typically 2–5 seconds. That's tight enough for most multi-venue investigations. The trade-off is complexity: you need a local server at each site anyway, plus a cloud subscription. You carry two failure points. However, you also dodge the bandwidth apocalypse of full cloud streaming. One blunt truth: hybrid forces your team to decide what “latency-sensitive” means before a crisis. Most managers guess wrong—they put a hallway camera on local and the parking lot on cloud.

“We spent six months building a hybrid handoff using VPN and AWS. We still lost three days of cross-site search because we forgot to set a shared time zone in the metadata schema.”

— Senior security architect, 14-venue retail chain

A rhetorical question worth asking: is your network team comfortable managing both edge storage rules and cloud ingestion policies simultaneously? Because the hybrid approach doesn't forgive amateurs. The metadata merger breaks if one site’s retention expires before the cloud batch job runs. The timeline merging—that clean cross-site view everyone wants—requires a shared ontology. Site A calls it “Zone 4,” site B calls it “Stockroom B,” and your search returns nothing. So the real choice is not cloud versus edge. The real choice is how much operational pain you tolerate to close that handoff blind spot. Right order: decide your latency budget, then pick the connection method. Wrong order: buy a platform, then discover your handoff architecture fights your existing cabling. Not yet convinced? The next section gives you five concrete criteria to kill the wrong options fast.

What Matters Most? Five Criteria to Judge Your Options

Latency: How Fast Can You Track a Person Crossing from Site A to B?

Most teams skip this: the gap between site A detecting a person and site B displaying that same data. I have seen setups where cloud upload adds three to seven seconds. That's an eternity when you're chasing a thief who knows the handoff zone. Worse — the person vanishes from one camera feed, then reappears on the next monitor only after a loading spinner. The catch is that cloud latency looks fine in a demo with one camera. Scale to fifteen venues and the backlog compounds. You want sub-second visibility across sites? Edge processing is your friend. But edge costs more upfront and the vendors rarely show you the real lag numbers until you sign.

Honestly — most physical posts skip this.

Honestly — most physical posts skip this.

Bandwidth Cost: Cloud Upload vs. Local Network Capacity

Here is where budgets quietly bleed. Uploading continuous 4K streams from every camera to the cloud for cross-site stitching burns through your monthly data cap faster than you expect. One client moved twenty cameras at two locations to cloud-only — their bill spiked by $1,400 in the first month. That hurt. The alternative: store locally and send only metadata or short clips across sites. But metadata handoffs require your venues to speak the same protocol. Most don't. What usually breaks first is the accounting team, not the network. They see the bill and freeze expansion. Worth flagging—some hybrid setups cut bandwidth by 60%, but only if you negotiate the codec and frame-rate drop *before* you deploy.

Redundancy: What Happens When the Internet Goes Down?

You lose WAN at site B. Does site A still record? Can security officers review yesterday's crossover footage locally, or does everything lock up? I fixed a system last year where the failure mode was total blindness — no internet meant no playback, no live view, no alerts. The local DVR was there but the software refused to serve anything without cloud verification. That's not redundancy; that's a single point of failure dressed in marketing. A proper handoff strategy stores a local buffer at each site, syncs when the line returns, and lets operators keep working offline. The trade-off: double the storage hardware and a more complex sync logic. But the alternative is a dark building when the fiber goes down. Not acceptable for a multi-venue operation.

Compliance: GDPR, HIPAA, and Retention Laws for Cross-Site Footage

Moving video between venues sounds simple until you cross a state line or a jurisdiction boundary. GDPR in Europe treats facial data as biometric — you can't just pipe a feed from Berlin to Madrid without explicit consent and retention limits. In the US, HIPAA applies if any site is a healthcare facility; one careless handoff that includes a patient's image in a hallway can trigger fines. The tricky bit is that most cloud vendors promise compliance but put the burden on *you* to configure retention policies per venue. I have seen teams forget site C's 30-day purge limit because site A uses 90 days. Cross-site footage then lingers illegally. Audit logs catch that, and the fines are not hypothetical — a hospital system I know paid $240K for exactly this oversight.

“We thought moving footage to the cloud solved compliance. It just moved the problem to a bigger, more expensive place.”

— Security director, three-venue retail chain, after a GDPR audit

Does Your Vendor Even Support the Real Scenario?

Fifth criterion, and the one that stings most when missed: can the software actually handle a person leaving site A, driving ten minutes, and re-entering at site B without manual timeline stitching? Many platforms advertise "cross-site tracking" but only work if both cameras are on the same LAN or same cloud tenant. Real handoffs involve different ISPs, different camera brands, and sometimes different timezone offsets. Test this with a live person walking the actual boundary — not a scripted simulation. If the vendor can't show you that working on your hardware, the blind spot exists. Don't sign until you see the seam close in real time.

Trade-Offs at a Glance: A Comparison Table

Table: Cloud vs. Edge vs. Hybrid — Latency, Cost, Redundancy, Scalability

The table below strips away the marketing gloss. I've built this from actual installs where the seam between sites either held or blew out completely. Each row maps to a real trade-off you will face — not a spec sheet fantasy.

Notice the latency column. That 3-to-8 second cloud gap is exactly where a person walks out of one venue's coverage, crosses a parking lot, and disappears before the next site's camera picks them up. That hurts. Edge cuts that gap to under a second — but you trade remote management ease. Hybrid tries to split the difference. The catch is that hybrid costs more upfront and demands someone who understands both network architectures, not just one.

Example scenario: 12 sites, 30 cameras each, 100 miles apart

Picture a regional retail chain: twelve locations scattered across three states, thirty cameras per store, and a security team that sits in one central office. We fixed this exact setup last year. Pure cloud meant every handoff dragged through a distant data center — fine for review, terrible for live tracking. Pure edge kept each store isolated; the central team lost visibility during shift changes. Hybrid won here: each site recorded locally for instant playback, while a lightweight stream fed the central NVR. The trade-off? Maintenance calls spiked because edge recorders needed quarterly disk swaps. Not a dealbreaker — but budget for it.

What usually breaks first in this scenario is the link. One store had a DSL line that died every Tuesday afternoon for three weeks. With a pure cloud system, that store went blind for hours. With hybrid, the local edge unit kept recording, and the cloud stream simply paused and caught up. Worth flagging — that catch-up process can spike your bandwidth bill if the outage runs long. I have seen bills jump $400 in a single month from that alone.

When edge wins: limited bandwidth, high uptime requirement

Edge shines in places where the internet is unreliable or painfully slow — think rural construction sites, remote warehouses, or seasonal pop-up venues. The decision flips based on one question: can you tolerate a two-second blind spot when a person crosses from camera A to camera B? If the answer is no — if you're tracking theft or safety incidents that unfold in seconds — edge is your only real option. The downside? You lose the ability to view live feeds from all sites on one screen without kludging a VPN together. That's a real operational hit. But for uptime, edge beats cloud every time — your system stays recording even when the ISP fails. Cloud-first systems return empty timelines during outages. Edge-first systems return footage. Pick your poison.

'The cloud looks great on a diagram. But when the fiber gets cut at 2 AM, edge is what keeps your video alive.'

— Lead integrator, 14-site retail deployment, 2023

When cloud wins: few sites, low latency needs, easy management

Cloud wins when your multi-venue footprint is small — say three office locations under one roof, each with ten cameras. If the handoff gap of 4–8 seconds is acceptable because nothing critical crosses between those buildings, the management simplicity of cloud is hard to beat. No local servers to patch. No disk arrays to replace. One login, one interface, one bill. That sounds fine until you add site number four — 40 miles away on a 15 Mbps uplink. Then the latency jumps again. The trap is that cloud scales up cost faster than complexity: per-camera fees, egress charges, retention over 30 days. I have watched a three-site cloud bill double when a client added five cameras per site and bumped retention to 60 days. Double. For the same hardware. So cloud fits best when you have few venues, stable internet, and zero tolerance for on-site hardware fiddling. Outside that sweet spot? The trade-offs start biting.

Flag this for physical: shortcuts cost a day.

Flag this for physical: shortcuts cost a day.

How to Implement Your Choice Without Breaking Things

Step 1: Network audit—measure latency and bandwidth between sites

You picked a strategy. Cloud, edge, hybrid—whatever. Now stop—don't touch a single camera feed until you know what your network actually does under load. Most teams skip this: they assume the WAN link between Venue A and Venue B can handle a simultaneous 4K stream plus access control traffic. That assumption breaks things fast. Run iperf between sites at peak hours. Measure jitter, not just raw bandwidth. The catch is that a 15ms ping looks fine on paper but kills frame-accurate handoffs when you need sub-second transfer of a suspect’s face across three buildings. I have seen a deployment stall for three weeks because nobody checked that the MPLS circuit had a 40% packet loss spike at 4 PM daily. Fix that before you wire anything.

What usually breaks first is the switch config. VLANs misaligned. QoS tags stripped. One concrete thing: set DSCP markings for video traffic to AF41—most routers respect that. If your IT team pushes back, show them the math: a single dropped frame at the handoff moment means a gap in coverage. That gap is a liability, not a network problem.

Step 2: Pilot at two venues with a test handoff scenario

Two sites. Not ten. Pick your busiest location and your quietest one. Stage a real handoff—someone walks a package from loading dock to lobby to parking lot. Track every second the feed blips or buffers. Pro tip: do this during low-traffic hours first, then repeat at shift change when chaos peaks. The difference will shock you. We fixed a client’s handoff by discovering that their cloud relay service dropped frames when the operator zoomed in—processing power bottleneck, not the network. A pilot catches hardware limits before you buy fifty units.

“The handoff looks clean in a demo. In production, it falls apart because nobody tested with the actual workload.”

— security ops lead, regional casino chain, after a botched rollout

Step 3: Train operators on the new workflow (not just software)

Here is where most implementations stumble. You roll out a new NVR interface, write a manual, call it done. Then the night shift operator at Site B misses the alert because they were trained on the old pop-up style, not the new tile view. Trained wrong is worse than untrained—they override the system. Run three scenarios: normal transfer, partial network failure (simulated), and a false alarm drill. Let operators click through the handoff button themselves until muscle memory kicks in. That said, avoid over-automation: one property manager insisted on full auto-tracking across sites, but the PTZ cameras kept locking onto reflections. Operators hated it, manually disengaged, and the handoff became a manual log entry—worse than before. Train judgment, not just clicks.

Step 4: Document a failover plan for network outages

The fiber goes down between sites. What happens? If your answer is “the cloud handles it,” test that. Most cloud handoffs buffer locally when disconnected—but the buffer size defaults to 30 seconds. A 45-second outage loses the critical handoff window. Document exactly: trigger thresholds (e.g., latency > 50ms auto-switches to local recording), fallback recording resolution (drop to 720p to save local storage), and manual escalation steps (who calls whom when the link stays down for 5 minutes). A pitfall to flag: don't rely on cellular failover unless you verified carrier coverage at the exact camera location. I visited a site where the 4G backup had one bar inside the loading dock—useless.

Patch test the failover. Pull the primary link during a drill. Watch operators fumble. Then fix the procedure. That hurts less than a real incident where you can't produce footage of the handoff zone.

Risks You'll Face If You Pick the Wrong Handoff Strategy

Delayed threat response: minutes lost stitching video together

That sounds fine in a spreadsheet. In reality, when a person of interest walks out of Venue A's parking lot and into Venue B's stairwell, the ten-second handoff lag becomes a ten-minute hunt. I have watched security teams burn forty minutes scrolling through two separate NVR interfaces, trying to match a hat colour and a limp. By the time they confirm the subject crossed the boundary, the car is already three blocks away. The math here is brutal: a twelve-second transfer delay—common with poorly configured cloud relays—costs you roughly a city block of coverage. Multiply that across five venues and you're no longer doing surveillance. You're doing forensics after the fact.

Gaps in evidence: missing timestamps or corrupted clips at merge points

Wrong order. Not just a packet problem—a legal problem. The tricky bit is that most handoff software timestamps each venue's feed independently. When two clocks drift by even three seconds, your "continuous" timeline snaps. Now picture this: a prosecutor asks why there is a two-second hole exactly when the suspect entered the corridor. You can't answer. Worse, some edge devices drop frames during the upload queue. You get a clip that jumps from 14:22:01 to 14:22:04. That seam is where people sue—and win. We fixed this for a client by forcing every camera to NTP-sync against the same stratum-2 server. Small fix, huge difference. But if you never test the merge points under load, you won't find the corruption until discovery.

“You don't notice the handoff is broken until you need that one second of footage—and it isn’t there.”

— former security director, three-hospital campus

Vendor lock-in: can't switch hardware without replacing all cameras

This is the trap that keeps on trapping. You choose a unified cloud platform that promises "any-camera compatibility." Six months later you realize it only works well with its own branded sensors. Now every new site needs the same proprietary edge box. Want to swap to a cheaper bullet cam from a different manufacturer? The handoff middleware refuses to trust its metadata stream. You're stuck. I have seen organisations pay quadruple for replacement cameras just to keep the handoff logic alive. The single smartest move is to ask one question before signing any contract: Can I mix brands without rewriting the integration? If the answer is "limited support," that's not limited—that's a future invoice.

Not every physical checklist earns its ink.

Not every physical checklist earns its ink.

Bandwidth surprises: cloud upload fees exceeding the cost of cameras

Most teams skip the bandwidth math. A 4K stream at 15 fps is roughly 6 Mbps per camera. Run five venues with eight cameras each—that's 240 Mbps continuous upload. Even with compression, you will chew through 2–3 TB per month. Now check your cloud provider's egress fees. The catch: they don't bill by camera; they bill by gigabyte pushed out to your VMS. One client discovered their monthly upload cost was $1,400 higher than the hardware lease for all sixty-two cameras. That's not a blind spot. That's the entire dashboard gone dark. Layer in burst pricing if a threat event triggers archive uploads, and the surprise bill hits before the investigation is closed.

Frequently Asked Questions About Multi-Venue Surveillance Handoffs

How much bandwidth do I need for 10 sites with 20 cameras each?

That number depends on three things: resolution, frame rate, and compression. But nobody gives you a straight answer because vendors want to sell you the fat pipe. Here is the dirty floor: 200 cameras at 1080p, 15 fps, H.265, will eat roughly 600 Mbps upstream under normal motion. That's not cheap. The catch—most sites share a single internet link with POS systems, Wi-Fi, and guest networks. I have watched a restaurant chain try to shove 20 cameras through a 50 Mbps upload. The seam between sites didn't just blur; it died. Plan for 40–50 Mbps per site as a safe minimum, then double it if you run analytics at the edge. Wrong order? You buy cloud storage you can't actually write to.

Can I use my existing analog cameras with a new cloud system?

Short answer: yes, with an encoder. Long answer: that encoder becomes your single point of noise. Analog cameras output a constant stream—no smart bandwidth shaping, no motion-triggered recording. You pay to upload static hallway feeds all night. That hurts. What usually breaks first is the handoff: the encoder lives at site A, but the cloud expects a clean RTSP stream. If the encoder buffers or drops frames, your cloud DVR shows gaps. The fix is a hybrid edge appliance that transcodes locally before sending—don't let the old coax dictate your cloud architecture. You will save money on cameras but spend twice on bandwidth.

“We kept our analog domes, added a $400 encoder per site, and our monthly cloud bill doubled in three months.”

— IT manager for a regional retail group, after a post-deployment audit

What happens to recorded video if the network goes down?

Most teams skip this:

Cloud-only systems store nothing locally. If a fiber cut or router failure hits between two venues during a critical handoff event—say a person leaving one parking lot and entering another across town—you lose that seam entirely. Not the footage from each site, but the continuity. Some vendors pitch “failover to SD card,” but I have yet to see a 256GB card survive 48 hours of 4K recording without corrupting. The smarter bet is a local NVR at each venue that caches 7 days and syncs once connectivity returns. That introduces a sync lag, however, and during that lag your central monitoring station sees a blind spot. Pick your poison: lose the handoff or risk stale video.

Do I need separate licenses for each site?

Depends on the VMS vendor. Some license per channel, others per server, a few per concurrent user. The pitfall: a “unified” license often means you pay per camera across all sites, even cameras that sit offline. We fixed this for a 15-venue client by auditing their license key count against active cameras—turns out they were paying for 300 cameras that had been replaced or decommissioned. Most vendors don't auto-revoke unused licenses. Ask for a true site-agnostic license pool. If the sales rep hesitates, you're about to buy a per-site lock-in that makes adding venue number 11 cost more than it should. That's the real blind spot—contract, not camera.

Final Recommendation: No Hype, Just What Works

For 2–5 nearby sites: cloud platform (e.g., Verkada)

If you run a handful of locations within reasonable latency of each other—think retail stores in one city, or small clinics across a metro area—a cloud-native platform like Verkada gets out of your way. I have seen this work well for three urgent-care clinics that share security monitoring during off-hours. Setup took an afternoon per site. The catch? Bandwidth. If any site has less than 10 Mbps upload, you will notice lag or dropped frames on playback. Fix that before you buy in. For five sites or fewer, the simplicity of one dashboard, one login, and auto-updating firmware beats stitching together local servers. But do not expect the same reliability during an ISP outage—cloud handoff dies when the pipe goes dark. That's the trade-off you accept for zero on-site server maintenance.

For 10+ sites with limited bandwidth: edge recording plus central dashboard

Most teams skip this: they assume cloud works everywhere. Wrong order. For a dozen warehouse distribution centers scattered across rural highways, cloud video becomes a slideshow. The smarter play is edge recording—store footage locally on each NVR, then push only metadata and alerts to a central dashboard. What break first is the metadata schema. If one site names cameras "Loading Bay 1" and another uses "Dock_A01," your search across venues collapses. I watched a logistics operator lose three hours tracing a theft because handoff logs used inconsistent labels. Spend the time upfront to standardize camera names, time zones, and event triggers across every site. That single step determines whether your "central view" is a searchable archive or a decorative thumbnail wall.

'We had bandwidth for two streams per site. Edge recording let us keep 90 days local and only send motion clips to the cloud. The central dashboard showed us what mattered—when, not constant video.'

— Security director, regional transportation company

For mixed environments: hybrid with local storage for critical cameras

The tricky bit is mixed environments—say, a corporate HQ with fat fiber and five remote kiosks on cellular backup. A pure cloud model suffocates the kiosks; pure edge leaves HQ blind to live feeds. Hybrid solves this: store all cameras locally at every site, but push a prioritized subset—the point-of-sale, the entry door, the server room—to a cloud aggregator in real time. That sounds balanced until you realize the human cost: you now have two systems to maintain. One concrete anecdote: a cinema chain I worked with installed hybrid across 14 locations. The cloud portion worked fine; the edge NVRs at two sites failed firmware updates silently, and nobody noticed for three weeks because the central dashboard still showed green status on the subset. That hurts. So budget for quarterly edge-NVR health checks, not just cloud dashboard uptime.

Bottom line: prioritize handoff metadata integrity over flashy features

Here is what I keep coming back to after auditing a dozen multi-venue deployments: the feature that looks sexiest—AI analytics, facial matching, license plate recognition—is worthless if the handoff drops the context between sites. A face match triggers at Site A at 14:03, but the video clip arrives at the central server with a timestamp shifted by +6 seconds because the cloud ingest queue stalled. You lose the chain of custody on that event. Most teams chase resolution and frame rates first. They should chase metadata integrity: camera naming conventions, atomic time sync (NTP on every recorder), and a schema that survives a network partition. That's not hype—it's the difference between a surveillance system you trust and one that produces evidence a lawyer can tear apart in cross-examination. Pick the strategy that preserves that seam, and everything else becomes a manageable upgrade.